PCA Request Line

[SABR Matt]

You used AJ Burnett, SavoyBG...I think that's who he meant.

[Matthew C.]

You used AJ Burnett, SavoyBG...I think that's who he meant.

Burnett, Beckett - whats the difference? :D

[SABR Matt]

Indeed. And incidentally, it doesn't show much of an appreciation for the game to completely ignore what's happening now in your analysis work, tune in to a game you say you don't even pay attention to (to the point where you can't name many current players by your own admission) and then start leaping to unfounded conclusions as though anyone with half a brain couldn't possible disagree when in fact there are many people (who have entire brains! :) ) who have done some of the necessary early work involved in proving the merits of looking at team defense as a context who would indeed disagree with you.

If you don't like baseball enough to pay attention to it today...why are you here?

[RuthMayBond]

Burnett, Beckett - whats the difference? :DRecession, repression, it's all the same thing

http://vids.myspace.com/index.cfm?fuseaction=vids.individual&VideoID=5017212#

[SavoyBG]

Indeed. And incidentally, it doesn't show much of an appreciation for the game to completely ignore what's happening now in your analysis work, tune in to a game you say you don't even pay attention to (to the point where you can't name many current players by your own admission) and then start leaping to unfounded conclusions as though anyone with half a brain couldn't possible disagree when in fact there are many people (who have entire brains! :) ) who have done some of the necessary early work involved in proving the merits of looking at team defense as a context who would indeed disagree with you.

If you don't like baseball enough to pay attention to it today...why are you here?


Are you saying that it's not possible to be very passionate about baseball unless you are very passionate about current major league baseball?

There's nothing wrong with looking at team defense. There's a lot wrong with taking the next step and saying that team defense is more important than pitching when it comes to balls in play. You guys are being fooled into thinking this stuff works because it seems to work with most average or near average pitchers, which covers a large percentage of major league innings. The system will not work well with the best and worst pitchers. If you went out there and pitched your batting average on balls in play would not be determined mainly by your fielders. It would be determined mainly by the whiplash you would have from turning to watch all the rocket line drives that you would be giving up. These numbers only SEEM to work because MOST major league pitchers are fairly close in ability.

As for people with brains agreeing, throughout history there have been many people with brains who disagreed with lots of things. For instance, most of the brightest people in the country in the late 1700s were sure that slavery was a necessary and justified aspect of life.

Besides, while there may be some agreement on this among the hard core Sabrmetric community, if you asked ALL people who are big baseball fans, including the players, managers, writers, etc... a vast majority of them would not buy into DIPS.

Matt, your problem is that you've developed this elitist attitude as if you're a member of some exclusive club of thinkers who are sure that you guys are smarter than all of the baseball people who don't care about all of these numbers so much.

Meanwhile a team has a better chance of winning if they throw the right pitch to Valdimir Gurerreo than if they know what DIPS says about Phil Hughes.

I'd love to see some major league team let you manage a season. You'd piss off your players so much that they'd be losing on purpose just to get rid of you.

You'd watch some pitcher give up 4 straight rocket line drives in the 6th inning and you'd leave him in because you'd think that it was just a fucntion of random chance.

Matt, you're not really a baseball fan. You're a numbers geek. You're much more concerned with elaborate mathematical formulas than you are with curveballs and sliders, or secondary leads.

Have you ever been on a baseball field with good baseball players, in any capacity?

Do you even know what a cockshot is?

As much as you know about SABRmetric numbers, you are not even remotely close to knowing as much about BASEBALL as i do. There is just so much to know about baseball, aside from mathematical formulas, that nobody in the world even comes close to knowing everything that there is to know.

Would you be able to show a high school pitcher how to throw a slider?

Would you be able to show a young 2Bman how to get his footwork right for turning the double play?

You really need a large dose of humilty, Matt, because you really don't know all that much about baseball.

[leewileyfan]

Indeed. And incidentally, it doesn't show much of an appreciation for the game to completely ignore what's happening now in your analysis work,

Up until now, I've largely stayed out of this thread, trying to recognize it as the SABR Matt private fiefdom.

I read Savoy BG's posts; and he's repeatedly stated that:

1. His primary interest is in the historical, generational context of the game.


2. He is not a sabermetric swami searching for the crystal ball to project future performance. He is interested in accomplishment than prognostication.

3. I essential share the same foci; so I can identify a bit with where he's coming from.

4. You set the stage in the quoted phrase above for another unfounded attack on his "analysis work" [redundant, that]


tune in to a game you say you don't even pay attention to (to the point where you can't name many current players by your own admission)

He paid enough attention to what happened so that he could give a play-by-play of the bazooka-try going on in the Angel's first inning.


and then start leaping to unfounded conclusions as though anyone with half a brain couldn't possible disagree when in fact there are many people (who have entire brains! :) )

Soundslike Olympic skating. You fail to mentionONE single, solitary unfounded conclusion on his part. You disagree .......... and predictably turn to the brain ............ [entire brains] .... what on earth does one do with a partial brain, depend on the part, I guess. How can you disagree with an unidentified conclusion by countering with "brains, entire brains," an no context or single challenged fact.


who have done some of the necessary early work involved in proving the merits of looking at team defense as a context who would indeed disagree with you.

Those who did that "necessary early work involved in proving the merits of looking at team defense as a contect" did all that long before you were born.


If you don't like baseball enough to pay attention to it today...why are you here?

If you cannot understand people more focused on comparative historical contexts of the game, with emphasis on the more general expanse of game components, then why are you here?

If you hadn't noticed, the name of this forum is: Statistics, Analysis & Sabermetrics > PCA Request Line:

1. This suggesys that the first three elements > PCA

2. Request Line suggests you are open to questions and welcome the burden of intelligent response;

3. The thread title does not seem to define intelligent inquiry being measured by one's immediate, eidetic grasp of who plays for whom at the moment ..... it suggests that it invites curious, diverse, probing inquiry ~ not some litmus test imposed by the host.

[brett]

I think DIPS is best used to evaluate DEFENSES. Since teams are likely to have extreme pitcher variations cancel out, you can get a pretty good gauge at defense from DIPS and also looking at things like doubles.

If you can evaluate a defense, then I prefer to evaluate pitchers using ERA+ and adjusting it to their defense.

Or if you prefer, I like OPS+ allowed times OPS+ allowed with runners on base.

Since pitchers basically allow runs by a multiplicative (pythagorian) effect of letting guys hit with the bases empty x getting hit with runners on base, (with the special exception of home runs).

Here are "self" relative OPS+ allowed with runners on base for some pitchers:


Greg Maddux: 115
Nolan Ryan: 113
Mike Mussina: 111
Randy Johnson: 107
Bert Blylevin: 104
Tom Seaver: 103
Phil Neikro: 103
Roger Clemens: 103
Johan Santana: 95
Zack Greinke: 94

If we believe that players have control over this, then we should divide a player's defense adjusted ERA or ERA+ by this factor.

Actually thinking about it though, no matter how you get hit with guys on base, if you don't get hit with the bases empty you also get a zero.

Ryan's overall split is 91 with the bases empty and 113 with runner on base. The product of that is 102.83. Thinking it through, that might be the more appropriate modifier. So If Ryan is a 123 dividing that by 1.0283 would make him effectively 119.6.

I'll have to look at some real instances. It is clear to me that Ryan did worse than the all-time greats with runners on base, and it is clear to me that getting hit with runners on base has more influence on runs than getting hit without runners on base.

[SavoyBG]

Up until now, I've largely stayed out of this thread, trying to recognize it as the SABR Matt private fiefdom.

I read Savoy BG's posts; and he's repeatedly stated that:

1. His primary interest is in the historical, generational context of the game.


2. He is not a sabermetric swami searching for the crystal ball to project future performance. He is interested in accomplishment than prognostication.

3. I essential share the same foci; so I can identify a bit with where he's coming from.

4. You set the stage in the quoted phrase above for another unfounded attack on his "analysis work" [redundant, that]



He paid enough attention to what happened so that he could give a play-by-play of the bazooka-try going on in the Angel's first inning.



Soundslike Olympic skating. You fail to mentionONE single, solitary unfounded conclusion on his part. You disagree .......... and predictably turn to the brain ............ [entire brains] .... what on earth does one do with a partial brain, depend on the part, I guess. How can you disagree with an unidentified conclusion by countering with "brains, entire brains," an no context or single challenged fact.



Those who did that "necessary early work involved in proving the merits of looking at team defense as a contect" did all that long before you were born.



If you cannot understand people more focused on comparative historical contexts of the game, with emphasis on the more general expanse of game components, then why are you here?

If you hadn't noticed, the name of this forum is: Statistics, Analysis & Sabermetrics > PCA Request Line:

1. This suggesys that the first three elements > PCA

2. Request Line suggests you are open to questions and welcome the burden of intelligent response;

3. The thread title does not seem to define intelligent inquiry being measured by one's immediate, eidetic grasp of who plays for whom at the moment ..... it suggests that it invites curious, diverse, probing inquiry ~ not some litmus test imposed by the host.

Thank you. I guess I'm not the only one who gets these condescending vibes from Matt.

[SavoyBG]

I think DIPS is best used to evaluate DEFENSES. Since teams are likely to have extreme pitcher variations cancel out, you can get a pretty good gauge at defense from DIPS and also looking at things like doubles.

If you can evaluate a defense, then I prefer to evaluate pitchers using ERA+ and adjusting it to their defense.

Or if you prefer, I like OPS+ allowed times OPS+ allowed with runners on base.

Since pitchers basically allow runs by a multiplicative (pythagorian) effect of letting guys hit with the bases empty x getting hit with runners on base, (with the special exception of home runs).

Here are "self" relative OPS+ allowed with runners on base for some pitchers:


Greg Maddux: 115
Nolan Ryan: 113
Mike Mussina: 111
Randy Johnson: 107
Bert Blylevin: 104
Tom Seaver: 103
Phil Neikro: 103
Roger Clemens: 103
Johan Santana: 95
Zack Greinke: 94

If we believe that players have control over this, then we should divide a player's defense adjusted ERA or ERA+ by this factor.

Actually thinking about it though, no matter how you get hit with guys on base, if you don't get hit with the bases empty you also get a zero.

Ryan's overall split is 91 with the bases empty and 113 with runner on base. The product of that is 102.83. Thinking it through, that might be the more appropriate modifier. So If Ryan is a 123 dividing that by 1.0283 would make him effectively 119.6.

Thanks Brett, I agree that DIPS is a good tool to evalute the effectiveness of a team's overall fielding, although you can never really know how much of it is just the pitcher's doing a good job of not letting hitters hit the ball solidly that often.

Those ERA+ numbers illustrate what I was saying about some pitchers just being better (or worse) with runners on base, scoring position, etc...

I just can't see the sense of assessing value based solely on components when we know exactly how many runs and how many earned runs that each pitcher allowed.

Baseball is not a game of Strat-O-Matic where every pitch or play is totally random and unrelated to the last dice roll.

[brett]

If you look at the best pitchers in the best seasons, their OPS+ allowed with runners on base is clearly lower than their overall OPS+ allowed. Perhaps one could argue that that is the luck component that makes statistically anomalous seasons, but I know that some have argued that pitchers DO save their better stuff for MOB situations.

Greike allowed an 88 OPS+ with the bases empty and 30 with runners on base this year.

Pedro 2000 allowed a 25 with the bases empty and 15 with runners on base.

Pedro '99: 48/24
Maddux '94: 38/26
Maddux: '95: 31/24
Randy Johnson '02: 77/47
Clemens '97: 51/50
Clemens '05: 56/32
Guidry '78: 56/42
Gooden '85: 64/23
Gibson: '68: 48/48
Jenkins '71: 92/80
Koufax '66: 71/42


Ryan '81: 42/52
Ryan '87: 51/64

Just some interesting stats for me. These are literally the first 13 seasons I picked and only Ryan was worse with runners on base.

It also looks like the best pitchers are even better with runners on base during their prime's. Is it the "luck" that makes a season statistically great in terms of ERA+ or is it a skill or by design?

Perhaps some is due to theA complicating factor may be that hitters tend to do worse with runners on second. Another might be that pitchers get bailed out by relievers with runners on base sometimes (and Ryan had poor bullpens).

[PVNICK]

Do you think there may be any component of Ryan's poor control per se leading to 3/2 counts and the runners going, moreso than most of the other pitchers?

[SavoyBG]

Do you think there may be any component of Ryan's poor control per se leading to 3/2 counts and the runners going, moreso than most of the other pitchers?

Probably a bit, but managers would be unlikely to send runners with less than two outs because of the high probability of a strikeout on that 3-2 pitch.

Naturally DIPS systems like Matt's say that Ryan was much better than his ERA+ numbers because it doesn't properly measure all of the little things that Ryan was poor at.

[brett]

Probably a bit, but managers would be unlikely to send runners with less than two outs because of the high probability of a strikeout on that 3-2 pitch.

Naturally DIPS systems like Matt's say that Ryan was much better than his ERA+ numbers because it doesn't properly measure all of the little things that Ryan was poor at.

Matt does account for wild pitches, poor CS rate of his catchers etc. It also is true that Ryan had poor defenses, ballparks that don't give him the best light and poor bullpens. It doesn't (as I understand it) count slugging allowed against a pitcher except for home runs, and it doesn't account for pitchers giving up higher than normal BABIP or pitching worse with runners on base.

[SavoyBG]

Matt does account for wild pitches, poor CS rate of his catchers etc. It also is true that Ryan had poor defenses, ballparks that don't give him the best light and poor bullpens. It doesn't (as I understand it) count slugging allowed against a pitcher except for home runs, and it doesn't account for pitchers giving up higher than normal BABIP or pitching worse with runners on base.

Amazing how they can disregard doubles and triples allowed by individual pitchers but still think that the system does a better job than using actual earned runs.

[PVNICK]

Probably a bit, but managers would be unlikely to send runners with less than two outs because of the high probability of a strikeout on that 3-2 pitch.

Naturally DIPS systems like Matt's say that Ryan was much better than his ERA+ numbers because it doesn't properly measure all of the little things that Ryan was poor at.

And then there is the idea that the control issues would be a detriment to fielders being on their toes, so to aspeak which would exacerbabte their lack of stellarness to begin with.

[SavoyBG]

And then there is the idea that the control issues would be a detriment to fielders being on their toes, so to aspeak which would exacerbabte their lack of stellarness to begin with.

According to Matt and the other SABR guys with this stuff, all fielders perform at the same level no matter who is pitching or what the situation.

That's what happens when you have a bunch of math geeks who never really played deciding what's what.

[brett]

Amazing how they can disregard doubles and triples allowed by individual pitchers but still think that the system does a better job than using actual earned runs.

I asked a question maybe a year or so ago in this site: why is Ryan's OPS+ against him so great but his ERA+ is not.

Basically you can line up his top ERA+ seasons against those of Seaver, and Ryan loses every one of them down the line, like 150 to 126, 144 to 124 etc, but he topped Seaver in OPS+ by scores of like 60 to 75!.

Seaver was allowing 75 OPS+ and 140 ERA+ and Ryan 60 OPS+ and 124 ERA+.

What TangoTiger told me was the using runs was better than using individual events for pitchers, but using events was better than using runs for hitters.

Also that Ryan lost out because he got stolen on, and threw wild pitches and had poor defenses.

I still think that Matt does 2-3 things better than any other system such as team defense, ballpark and pitching staff tendancies and you can't overlook those things and expect your method to be any better.

I DO believe that pitchers control runs, not just events. I also feel that a pitcher can get hit hard even though he does all the DIPS things well. I've watched Ubaldo Jimenez for the last few years and the guy has nearly unhittable stuff for 5 or 6 innings, but he will have 2 rough innings every time. He will string "bad things" together.

[SavoyBG]

What TangoTiger told me was the using runs was better than using individual events for pitchers, but using events was better than using runs for hitters.



This is what I've been saying all along here, with Matt ridiculing me as if I just don't get it.

Maybe he'd like to also ridicule Tango, who has a lot more credibility than he does.

[SABR Matt]

I would agree with Brettt that evaluating team defense (the fielding chunk) must begin with DIPS. In fact, that's exactly how PCA uses DIPS (the pitchers then split the credit assigned to pitching using the DIPS divide) based on their contributions to the team's DIPS line and to the distribution of batted ball results as data allows.

And no...I'm not saying it's impossible to be passionate about baseball without being passionate about current baseball...I am, however, saying it's impossible to render accurate judgments about what you're seeing when you decide to randomly tune in to a current baseball game without any current baseball context. If you're not watching the current major leagues...you can't comment about the inner workings thereof. Not with any analytical skill.

[Matthew C.]

I DO believe that pitchers control runs, not just events. I also feel that a pitcher can get hit hard even though he does all the DIPS things well. I've watched Ubaldo Jimenez for the last few years and the guy has nearly unhittable stuff for 5 or 6 innings, but he will have 2 rough innings every time. He will string "bad things" together.

My understanding is that PCA initially did not account for XBH prevention, but when he created DNRA+ from PCA a few years back, he did include XBH prevention. I am sure Matt will answer this.

Also - did Tom say this considering career runs, or seasonal runs? I have no doubt that he said this regarding career runs allowed, but I have a hard time believing this would come out of the mouth of the creator of FIP regarding season-by-season RA.

Finally, Matt has never denied that pitchers do have some control over BABIP, nor has he denied that pitchers can string together a bunch of legit hard hit balls, nor does DNRA+. Not sure why everybody is still alluding to Matt holding a McCracken viewpoint.

Also Savoy -I am confused with the "assumes all fielders perform at the same level..." statement. It is ERA and ERA+ that assumes equal Fielding for every pitcher, most saber metrics try to adjust for different fileding abilities - that is one of their main functions.

PVNICK - no study has shown anything to the "control issues correlated to fielders attentiveness/quality idea. The poor fielders behind Ryan played poorly behind every pitcher they played with.

[SavoyBG]

And no...I'm not saying it's impossible to be passionate about baseball without being passionate about current baseball...I am, however, saying it's impossible to render accurate judgments about what you're seeing when you decide to randomly tune in to a current baseball game without any current baseball context. If you're not watching the current major leagues...you can't comment about the inner workings thereof. Not with any analytical skill.


Baseball is baseball and some points can be made with any era as an illustration of the point.

You don't need any analytical skill to see when a pitcher is getting hammered as opposed to when he's pitching well and the batters are getting a few hits here and there and not really hitting the ball solidly.

Let's say there's about 400 pitchers in the major leagues. About 300 of them will seem to be accurately guaged by DIPS. The problem is that there are about 50 pitchers who are a cut (or more) above the other 300, and about 50 pitchers who are a cut (or more) below the other 300, and DIPS will not do a good job of evaluating those pitchers.

The Seaver / Ryan stuff is a perfect example. DIPS methods would show Ryan as having as good or better value than Seaver, but in reality, Seaver was not only better, but a lot better.

During Ryan's prime we all wondered why this guy who seemed so unhittable would give up more runs than it seemed he should based on his hits and walks and HRs allowed. When it goes on year after year it's not luck anymore. He was considerably worse from thye stretch, he couldn't hold runners on at all, wild pitches, his own bad fielding, etc... all made him worse than DIPS oriented methods would show now.

Nobody in the 70s thought that Ryan was as good as Seaver, or Palmer, or Carlton, and he wasn't. At his best on sveral particular days he was better than any of them, but not for a season or for a career.

[leewileyfan]

And no...I'm not saying it's impossible to be passionate about baseball without being passionate about current baseball...I am, however, saying it's impossible to render accurate judgments about what you're seeing when you decide to randomly tune in to a current baseball game without any current baseball context.

Do you ever read this stuff before you type it?

Is there no difference between not "being passionate about current baseball" and the other extreme of "randomly tuni[ng] in to a current baseball game?"

If one, like me, becomes a young fan @ 1939; learns the game and plays the game a lot, say through entry into the Army as a draftee, becoming increasingly avid along the way, then gradually becomes a bit less passionate about the current game, favoring historical context:

1. Who are you to diagnose a case of out-of-touch-out-of-mind amnesia?

2. When one learns the game over a long time, plays it over a considerable stretch, he may not be avidly devoted as he/she once was to each game-day; that does not disqualify the person as a knowledgeable fan. It's like saying somebody's grandmother, who cut school to see Frank Sinatra at the Paramount [all decked out in bobby-socks and saddle shoes or penny loafers] can no longer appreciate or comprehend music.


If you're not watching the current major leagues...you can't comment about the inner workings thereof. Not with any analytical skill.

Yeah, that qualifies for #3. Gotta ove "the inner workings thereof." Reminds me of A Clockwork Orange.

Humble up. Get with the program ..... any program!

[SavoyBG]

Also Savoy -I am confused with the "assumes all fielders perform at the same level..." statement. It is ERA and ERA+ that assumes equal Fielding for every pitcher, most saber metrics try to adjust for different fileding abilities - that is one of their main functions.



What I mean is that the players are human beings and there are other factors that could make players play better or worse behind a certain pitcher.

Take a look at how the Phillies played in 1972 when Carlton was pitching as opposed to how they played when he wasn't pitching.

If players believe that they have a good chance to win they are going to play better than if they think there's no hope.

They may play worse if they get sick of a pitcher who can't throw strikes.

I've been on baseball fields working games when things like this have occurred.

[brett]

My understanding is that PCA initially did not account for XBH prevention, but when he created DNRA+ from PCA a few years back, he did include XBH prevention. I am sure Matt will answer this.

Also - did Tom say this considering career runs, or seasonal runs? I have no doubt that he said this regarding career runs allowed, but I have a hard time believing this would come out of the mouth of the creator of FIP regarding season-by-season RA.


Career. Post #5
http://baseball-fever.com/showthread.php?t=74461&highlight=seaver+ryan

[SavoyBG]

Do you ever read this stuff before you type it?

Is there no difference between not "being passionate about current baseball" and the other extreme of "randomly tuni[ng] in to a current baseball game?"

If one, like me, becomes a young fan @ 1939; learns the game and plays the game a lot, say through entry into the Army as a draftee, becoming increasingly avid along the way, then gradually becomes a bit less passionate about the current game, favoring historical context:

1. Who are you to diagnose a case of out-of-touch-out-of-mind amnesia?

2. When one learns the game over a long time, plays it over a considerable stretch, he may not be avidly devoted as he/she once was to each game-day; that does not disqualify the person as a knowledgeable fan. It's like saying somebody's grandmother, who cut school to see Frank Sinatra at the Paramount [all decked out in bobby-socks and saddle shoes or penny loafers] can no longer appreciate or comprehend music.



Yeah, that qualifies for #3. Gotta ove "the inner workings thereof." Reminds me of A Clockwork Orange.

Humble up. Get with the program ..... any program!

The funniest thing is that he actually believes that because he is a numbers geek that he is hip to the "inner workings" of major league baseball.

This is a guy who I would bet could not even throw a ball from the plate to 2B on a fly. I bet I could strike him out easily, and my top fast ball is about 65 MPH.

I'd bet that he could not look at a hitter in the cage and have any idea what he was doing wrong or right on any particular swing.

Matt, what would you do if you were named the coach of a baseball team with no stats involved?

What would you do on the first day of practice?

How would you figure out your batting order without numbers to go on?

Matt, you've a very good mathemetician, but you're a poor judge of your status if you think that you are some kind of baseball insider.

[brett]

I would agree with Brettt that evaluating team defense (the fielding chunk) must begin with DIPS. In fact, that's exactly how PCA uses DIPS (the pitchers then split the credit assigned to pitching using the DIPS divide) based on their contributions to the team's DIPS line and to the distribution of batted ball results as data allows.

And no...I'm not saying it's impossible to be passionate about baseball without being passionate about current baseball...I am, however, saying it's impossible to render accurate judgments about what you're seeing when you decide to randomly tune in to a current baseball game without any current baseball context. If you're not watching the current major leagues...you can't comment about the inner workings thereof. Not with any analytical skill.

Matt, whether it is due to skill or luck, what affect would Ryan having roughly a 113 "self relative" OPS+ with runners on base do to his ERA+?

Keeping it simple, what it a team batted .300 but .320 with runners on base, or a pitcher gave up .300 but .320 with runners on base (about 6.5% given that it gets doubled in the OPS+).

[brett]

What I mean is that the players are human beings and there are other factors that could make players play better or worse behind a certain pitcher.

Take a look at how the Phillies played in 1972 when Carlton was pitching as opposed to how they played when he wasn't pitching.

If players believe that they have a good chance to win they are going to play better than if they think there's no hope.



I've wondered about this. Definitely they scored more. In fact if memory serves correct from the stats, The Phils averaged about 3.00 when he didn't pitch and 4.00 when he did.

[leewileyfan]

I could just as easily point to any of a number of starts in which the pitcher literally gave up 4, 5, 6 bloop and dink hits in a row and say "how could anyone watching this game believe that ball in play hits are the pitcher's fault? Everything these guys hit is a duck snort or a 40 hopper squeaking into the outfield!"

I remember a start made by Felix Hernandez back in the first couple of months of 2006 or 2007 in which he gave up a sequence of hits that went something like:

Infield single to third
Bunt Single to the catcher
Sac Bunt
Bloop single over second base (2 runs score)
Infield single to short
Seeing eye single through the third base hole (1 run scores)
K
infield single to first, loading the bases
Monstrous grand slam home run - Felix out of the game.

Are you really going to sit here and tell me all of those hits were Felix pitching poorly?

Getting back to the specific post that I believe started these exchanges [quoted above], here's how I see that series:

1. Batter swung for something more threatening; but made weak contact ~ a big break for the hitter = LUCK

2. "Bunt single" [to the catcher is irrelevant] .... batter attempted bunt contact and succeeded ..... score one for the offense [even if you could as easily put some blame on the catcher, the pitcher, or someone else on defense];

3. "Bloop single over 2B" .... well we do know that the batter swung away, made contact and hit the ball at least 130' or more beyond the range of anyone in the defense:

-that's either lousy range at SS or 2B; OR, a bloop hit with just enough authority and speed in flight to let the fielders off the hook;

4. Infield single to short ... no other defensive comment suggests it was reasonably well hit and presumably deep enough in the hole to be scored a hit;

5. "Seeing eye single through the third base hole" .... sounds like a ball hit fast enough to evade even a decent range from either SS or 3B;

6. K ~ Since this is it might well be the opposing pitcher in the NL or the 8th/9th batter in the AL; since the MLB K rate is @ 17-19%, no great shakes getting one by now;

7. Infield hit to first ............................

Anyone who has a working knowledge of the game has, by now, realized these guys are picking up the ball well, perhaps swinging too eagerly and thus hurting themselves more than the pitcher is fooling them. A trip to the mound by someone would have been called for, probably before, ceratinly after, the K.

8. Monstrous grand slam. A savvy coach or manager [I]might have seen the writing on the wall.

Dinks, bloops, bunts successfully executed and humpback liners often hint at trouble 'round the bend.

[Matthew C.]

Career. Post #5
http://baseball-fever.com/showthread.php?t=74461&highlight=seaver+ryan

I am about 100% sure that Matt would agree with this.

[brett]

I am about 100% sure that Matt would agree with this.

Including this part?

Ryan was terrible with men on base, compared to himself with bases empty:
http://www.baseball-reference.com/pi...62=1&lg=&team=

[Matthew C.]

Including this part?

Ryan was terrible with men on base, compared to himself with bases empty:
http://www.baseball-reference.com/pi...62=1&lg=&team=

No - the part about career RA holding up better on their own than seasonal RA.

[SavoyBG]

Including this part?

Ryan was terrible with men on base, compared to himself with bases empty:
http://www.baseball-reference.com/pi...62=1&lg=&team=

What is there not to agree with?

Either he was much worse with men on base than with the bases empty or he wasn't. It's not a matter of opinion.

Maybe there's a little wiggle room for defining "terrible" as opposed to just "bad," but the numbers are the numbers.

[Matthew C.]

What is there not to agree with?

Either he was much worse with men on base than with the bases empty or he wasn't. It's not a matter of opinion.

Maybe there's a little wiggle room for defining "terrible" as opposed to just "bad," but the numbers are the numbers.

That has always been Matt's contention - that compared to the average pitcher, Ryan was just slightly bad with m.o.b. He always looks at pitcher compared to average, not pitcher compared to other HOFers. I have had many converstion with Matt about situational splits, and he has said that he intends on investigating more and including any significant results in his rankings.

[SavoyBG]

That has always been Matt's contention - that compared to the average pitcher, Ryan was just slightly bad with m.o.b. He always looks at pitcher compared to average, not pitcher compared to other HOFers. I have had many converstion with Matt about situational splits, and he has said that he intends on investigating more and including any significant results in his rankings.


Okay, so if he's an all time great with nobody on base but slightly bad with runners on base than you have to factor that into his rating.

Apparently Matt doesn't do that because I know his numbers have Ryan ranked much better than he is ranked in ERA+ based systems.

[Matthew C.]

Okay, so if he's an all time great with nobody on base but slightly bad with runners on base than you have to factor that into his rating.

Apparently Matt doesn't do that because I know his numbers have Ryan ranked much better than he is ranked in ERA+ based systems.

I agree that Matt needs to consider this more, but the biggest reasons that Ryan's DNRA+ is much better than his ERA+ are:

1. Ryan's poor defenses cost him 3-4 ERA+ points.
2. Ryan's horrible bullpens cost him another 1-2 ERA+ points due to allowing a much worse-than-average number of inherited runners to score.
3. Ryan's BR park factors affect him too much (the large size of his parks matter less to Ryan, considering he allowed much less contact and many fewer FB than the average pitcher. BR park factors assume all pitchers are affected the same by the park) Here is another 2-3 ERA+ points.

ERA+ has Ryan at about 111, DNRA+ has Ryan at 123 (I think). I estimate that when all things are considered (the 3 points above and situation pitching) that Ryan is around 118-119 in reality. Matt does need to include the splits, in my opinion, but Matt, factoring in the 3 points above but not situational splits will get you closer to Ryan's real "ability" than ERA+ including splits but ignoring the 3 above.

[BigRon]

That's what happens when you have a bunch of math geeks who never really played deciding what's what.

I do believe that there is a disconnect between some of the newer math- analytical types and some who grew up with and were heavily involved in the human side of the game. That said, I strongly believe that a lot has been "uncovered" about baseball in the past 20 years or so, much- not all- of it uncovered by the new breed of analysts. I also believe that there may be an overreliance by some on "math" to understand the essence of baseball.

That is one reason why I have always been, and continue to be, a huge fan of Bill James. In my opinion he has asked more good questions in the last 30 years, and answered many of them well, than anyone else involved in baseball. James has a belief in analysis, but he is not a slave to it. That is why there is an "emotional" element to his analyses and observations, and that probably is why he frustrates some of the harder-core analytical types.

Good analytical techniques and conclusiions are valuable. So is reliance on physical knowledge and experience. Those who combine the two sides well are those who do the best job of understanding the game.

[SavoyBG]

I do believe that there is a disconnect between some of the newer math- analytical types and some who grew up with and were heavily involved in the human side of the game. That said, I strongly believe that a lot has been "uncovered" about baseball in the past 20 years or so, much- not all- of it uncovered by the new breed of analysts. I also believe that there may be an overreliance by some on "math" to understand the essence of baseball.

That is one reason why I have always been, and continue to be, a huge fan of Bill James. In my opinion he has asked more good questions in the last 30 years, and answered many of them well, than anyone else involved in baseball. James has a belief in analysis, but he is not a slave to it. That is why there is an "emotional" element to his analyses and observations, and that probably is why he frustrates some of the harder-core analytical types.

Good analytical techniques and conclusiions are valuable. So is reliance on physical knowledge and experience. Those who combine the two sides well are those who do the best job of understanding the game.

I agree Ron. I'm more into the numbers by far than anybody who I know in person, and was like that before I ever heard of Bill James. I was using my own crude formulas as far back as the early 1970s when I was a young teenager.

What brought me back some to other aspects aside from the numbers was when i started umpiring back in 1996. Over those years I learned a lot about the game that I never knew as a fan or a stats guy. I thought I klnew the rules until I took the class to become an umpire.

Not that I've ever been on the field with anything approaching major league level players, but I have done some of the best men's ameteur teams in the country as well as some decent lower level college ball. Many of the players I've worked with were former pro players including a few former major leaguers like Jeff Bittiger and some kids in high school who later became major leaguers like Vin Mizarro. I had the plate for two or three of his games.

Like I said before, and this is actually a Bill James quote that he made during some argument we had years ago on a message board. Not sure if this was exactly what he said, but it was to this effect....

"Nobody really knows all that much about baseball because there's just so much to know."

He's right. There's loads of coaches and players who know how to play baseball and how to teach others about baseball, but don't know hardly anything about the rules.

There's loads of historians who don't know the first thing about how to throw a slider.

And there's loads of mathemetician stat guys who don't know very much about baseball aside from the numbers.

There's just too much to know.

[brett]

There's loads of historians who don't know the first thing about how to throw a slider.

And there's loads of mathemetician stat guys who don't know very much about baseball aside from the numbers.

There's just too much to know.

That may be true, but I think that major league managers have major gaps due to failure to understand the stats of scoring. They still regularly bunt a guy to second with one out even with a top of the order hitter-not a great hitter but a typical #2 guy for example.

They typically do not tend to put higher OB% guys in the lineup in favor of higher BA guys who produce less.

Take for example the Rockies with Yorvit versus Ianetta at catcher. Ianetta produced much more because he walked and had power even with a .228 BA, but Yorvit got time with his .290 BA. He appeared to produce but the Rockies considered Ianetta to have been a failure this year with a 100+ OPS+ at catcher.

[SavoyBG]

That may be true, but I think that major league managers have major gaps due to failure to understand the stats of scoring. They still regularly bunt a guy to second with one out even with a top of the order hitter-not a great hitter but a typical #2 guy for example.

They typically do not tend to put higher OB% guys in the lineup in favor of higher BA guys who produce less.

Take for example the Rockies with Yorvit versus Ianetta at catcher. Ianetta produced much more because he walked and had power even with a .228 BA, but Yorvit got time with his .290 BA. He appeared to produce but the Rockies considered Ianetta to have been a failure this year with a 100+ OPS+ at catcher.

I haven't seen any position players sacraficing with one out.

Yes, there are definitely managers who do some stupid things. Frank Robinson with the Nationals was bunting in the first inning like crazy.

[leewileyfan]

What prompted me to post this was the discussion on Nolan Ryan. The following is my listing of top starters from 1901 on, who are listed in order of ERA, adjusted for batting climate and generation, WHIP elements and relative dominance. It is an approach that would appear to favor outstanding power guys; but some finesse pitchers come in surprisingly well.

To qualify, a pitcher must have had at least 100 ML starts [most have far, far more]; and this was to accommodate pitchers whose careers were shortened by war or death or illness. My intent in creating the list was to consider outstanding effectiveness reasonably disregarding highest volumes of IP. which would only regurgitate old-hat listings.

No active pitchers are included; and you may point out some omissions. Some HoF pitchers are not listed because they did not get a rating at or above the lowest listed.

Some guys listed will never see the HoF; and I am not picketing for their inclusion. I am merely recognizing brilliance on the mound for guys with late starts, military call-up, sudden injury, illness, death ... a few with more than one of the above. None of these guys was ever considered to be "a flash in the pan."

It's career ERA, as adjusted to 1901-1999 overall [inactives] against a standard for that population, like an ERA+:

W. Johnson 1.872
Grove 1.832
Hubbell 1.798
Walsh 1.744
Clemens 1.660
Maddux 1.652
Joss 1.650
Vance 1.532
Mathewson 1.519
Seaver 1.509
Young 1.503
Alexander 1.500
Chandler 1.482
Hughson 1.479
Gomez 1.448
Warneke 1.416
Brecheen 1.403
Schilling 1.395
Feller 1.395
M. Cooper 1.390
Mussina 1.388
Koufax 1.367
Whitey Ford 1.365
M. Brown 1.351
Dutch Leonard 1.342
Spahn 1.297
Derringer 1.295
Walters 1.284
M. Lanier 1.279
Lyons 1.277
C. Mays 1.275
Maglie 1.267
Waddell 1.266
Eckersley 1.250
T. Bridges 1.244
Ruffing 1.240
J. Allen 1.228
Plank 1.225
R. Ford 1.224
Palmer 1.221
Marichal 1.218
Glavine 1.207
Cuellar 1.206
Messersmith 1.206
Colon 1.204
Ryan 1.200
Guidry 1.198
Gibson 1.182
Newcombe 1.180
Newhouser 1.168
Bunning 1.163
Pierce 1.161

[Domenic]

Could you provide a comparison of Mike Piazza, Jorge Posada, and Ivan Rodriguez? I'm particularly interested in defense, but the offensive side of the equation is interesting, as well, as Rodriguez has lesser numbers, but a great deal more plate appearances.

Thanks in advance.

[leewileyfan]

Domenic: Since you didn't address your question on the three catchers to anyone in particular, I'll ytake a short at it [that way you'll probably get more input than you wanted]:

Offensively, they line up like this:

1. Piazza Runs Created per Plate Appearance = .1706; so in 600 PA, Piazza [career] averaged @ 102.36 RC/600 PA;

2, Jorge Posada, similarly calculated = .1512 or 91.27 RC/600 PA;

3. Ivan Rodriguez = .1362 0r 81.73 RC/600 PA

Defensively, listed by my ratings, vs. MLB catcher averages, and converted to Defense Runs:

1. I-Rod = .973 > .940 = 241 defense runs, career

2. Posada = .938 < .940 = -11 defense runs, career

3. Piazza = .929 < .940 = -57 defense runs, career

If we take overall offensive runs, modified up or down solely by defense runs as catcher, we get something like:

Net runs:

1. Piazza = .1632

2. Posada = .1504

3. Rodriguez = .1610

Another approach, limiting batting runs to only those while catching [I do not do this] would give a clear overall position edge to I-Rod.

[Domenic]

Thanks, Lee. I was looking for PCA data (hence my posting in the PCA Request Line), but I do appreciate any and all input. There is a thread on this very topic in the Hall of Fame sub-forum, and I was looking for some data/input beyond BP and WAR.

[SABR Matt]

I am about 100% sure that Matt would agree with this.

DNRA+ suggests that most of the time, career ERA is a better guess than career FIP or bare-bones DIPS ERA. My position is that we can do both things at the same time very well with a little attention to detail. We can account for the luck factors and the defense (as a context) that screw up single-season run prevention statistics AND we can create career evaluation tools that come close to replicating the power of ERA without any of the potential career-level biases (some pitchers really do spend most of their careers pitching in front of better than average defenses, some pitchers really are worse than average at BABIP prevention (or better), some really do have consistent trends with runners on base, etc. These things need to be analyzed and ERA+ even at the career level still has POTENTIAL to mislead the user...potential I can significantly reduce.

[SABR Matt]

That has always been Matt's contention - that compared to the average pitcher, Ryan was just slightly bad with m.o.b. He always looks at pitcher compared to average, not pitcher compared to other HOFers. I have had many converstion with Matt about situational splits, and he has said that he intends on investigating more and including any significant results in his rankings.

This is correct. Ryan was...a little worse than average...with men on base...but we're not talking about a huge gulf. We're talking about what may amount ot on the order of one or two runs per year (a handful of ERA+ points at best).

And I certainly do agree that a split-based correction needs to be applied to future incarnations of DNRA. The reason I have hesitated on this in the past is that I would like to determine a way to account for real changes in performance with M.O.B. that does not introduce the potential for major single-season errors caused by simple bad luck with men on base. I would point out that even for pitchers who have an overall career split that diverges from the norm with M.O.B. like Ryan, I can point to wild swings in M.O.B. performance that have absolutely nothing to do with the pitcher and everything to do with random clustering of hits.

I think the key will be looking at DEFENSE INDEPENDENT splits with men on base. It does not matter what type of pitcher you are or what kind of luck you get...if you change your approach in a way that really does influence the game over the long term, it will change your DIPS categories. When Glavine gets guys on base, his walk rate goes up, his K rate goes up, and his HR rate goes way down. Why? Because with people on base, he gets a lot more careful about his location and nibbles against guys who are more likely to hit him hard. So his career walk rate is probably misleading because many of those walks will have been with a base empty and thus unleveraged significantly compared to the average run-value of a walk.

I will need to think about how best to do this, but that's where I am right now.

[SavoyBG]

Saunders hasn't allowed a run yet after three innings, but it's time to get him out of there. With 6 baserunners already Scoscia just checked his PCA rating for the game and it's not very good. With a WHIP of 2.00 he'as actually pitching very badly and should be replaced now.....right Matt?

[SavoyBG]

Saunders hasn't allowed a run yet after three innings, but it's time to get him out of there. With 6 baserunners already Scoscia just checked his PCA rating for the game and it's not very good. With a WHIP of 2.00 he'as actually pitching very badly and should be replaced now.....right Matt?

See what I mean?

[Second Base Coach]

The phone to the bullpen must only be outgoing... no way for Matt to dial in...

[brett]

I think the key will be looking at DEFENSE INDEPENDENT splits with men on base.

This might be avoidable. If we can get a league DNRA with the bases empty and with men on base we could probably assume that all defenses are affected in generally the same way with men on base.

Ex if the league has a 101 with the bases empty and 97 with men on, that ratio.

By the way, I was thinking that if a guy like Ryan is doing worse with runners on, then he is doing better at keeping a guy off base in the first place.

If we take an average guy, and make him virtually unhittable (DIPS) in the MOB split then he's not going to give up runs except for home runs, BUT if we take an average guy and make him virtually unhittable with the bases empty he's not going to give up many runs either.

So its not a matter that having a better split with MOB will always be better.

The product of the two splits will yield a PEAK runs allowed number at some point very close to 100/100 split (it has to be because the 0/200 and 200/0 splits are going to yield zeros). It's not right at the 100/100 relative production level because if you get hit with men on base, you still put a guy on base.

[SABR Matt]

No...it's not that I'm imagining that defenses are getting worse or better with people on base differently for different people...I am saying that from year to year, there will be wild fluctuations in that split for an individual pitcher based on the random variation of his hits. If you look at Ryan, for example...his MOB/BE split varies from 0.90 to 1.30...it just averages out to 1.13. The years where he happened to allow a 0.90 ratio (lower OPS with MOB than with BE) didn't get that way because he was legitimately doing something differently that enabled him to lock down the running game. It was random chance...luck...whatever you want to call it.

[SABR Matt]

SavoyBG just did a spectacular job proving himself ill informed on this issue, BTW. Not to belabor this point, but Pettitte was clearly outpitching Saunders from the DIPS perspective...the difference was obvious by the end of the second inning. The fact that the score went to 1-0 by the end of the third didn't phase me...I knew Saunders was pitching a good notch worse and the Yankees did indeed eventually get to Saunders while Pettitte was busily pitching a gem.

The sad thing is...I don't think Savoy will ever see why I could tell the difference. I doubt he's ever going to understand the value of missing bats without walking guys as an indicator of pitching skill that translates to sustained success and real value.

[SavoyBG]

No...it's not that I'm imagining that defenses are getting worse or better with people on base differently for different people...I am saying that from year to year, there will be wild fluctuations in that split for an individual pitcher based on the random variation of his hits. If you look at Ryan, for example...his MOB/BE split varies from 0.90 to 1.30...it just averages out to 1.13. The years where he happened to allow a 0.90 ratio (lower OPS with MOB than with BE) didn't get that way because he was legitimately doing something differently that enabled him to lock down the running game. It was random chance...luck...whatever you want to call it.

How do you know that he wasn't doing something differently, or that the catcher wasn't?

Is it random chance when a batter like david Wright all of a sudden stops homering for a season?

[brett]

No...it's not that I'm imagining that defenses are getting worse or better with people on base differently for different people...I am saying that from year to year, there will be wild fluctuations in that split for an individual pitcher based on the random variation of his hits. If you look at Ryan, for example...his MOB/BE split varies from 0.90 to 1.30...it just averages out to 1.13. The years where he happened to allow a 0.90 ratio (lower OPS with MOB than with BE) didn't get that way because he was legitimately doing something differently that enabled him to lock down the running game. It was random chance...luck...whatever you want to call it.

Perhaps, but look at guys other than Ryan. All of the great seasons I listed before had MOB splits that were better than their overall OPS+ allowed rate.

What I felt-not a statistical argument-was that top pitchers tended to be better with MOB during their primes than during their early or decline periods.

At the very least, the best seasons seem to have great MOB OPS+ splits. Perhaps that is the "luck" that makes a season stand out, but it happened in basically every great year I could find, but NOT in Ryan's best years.

I also realized that the Empty/MOB splits would have different effects in different offensive settings. In some settings, the optimal spit would be different than in other settings. Also we have to account for IBB's and also steals don't factor into OPS+ but do affect run scoring.

Anyway, I think its important though confounding. I may be seeing the math wrong, but the way I see it, Ryan's 113 OPS+ allowed relative split with MOB creates as much as a 10 point uncertainty in a player's career DNRA+.

The first thing I'd be interested in is seeing how the splits actually effect runs allowed for real pitchers-The ratio of runs allowed to a DIPS score correlated to various MOB splits.

If we knew how the MOB split affects overall runs/DIPS for a very large number of players, all of the other factors should wash out.

In fact we might want to use teams as our subjects, or even leagues.

So let's say that for every team in history that we find that:

Runs allowed/DIPS expected runs allowed = M*MOB split + B

we have a formula to solve the problem.

[SABR Matt]

Umm...

Duh?

Of course great seasons typically come with good performance with men on base. The ERA tends be lower when you...um...don't give up hits with guys on base. Yep. Check.

That doesn't mean the year to year fluctuations in performance in that category are anything more than random. Think about this, Brett...when I approach a question like this, I start off by going to first principles...what physical explanations could there be for sustainable skill in preventing baserunners from scoring better than average? Other than the pitcher himself being intelligent in his plan when he gets guys on base, there is nothing to explain variations in M.O.B. performance. So unless yo're arguing that pitchers get dumber and smarter from year to year...single-season changes are meaningless.

[Matthew C.]

Umm...

Duh?

Of course great seasons typically come with good performance with men on base. The ERA tends be lower when you...um...don't give up hits with guys on base. Yep. Check.

That doesn't mean the year to year fluctuations in performance in that category are anything more than random. Think about this, Brett...when I approach a question like this, I start off by going to first principles...what physical explanations could there be for sustainable skill in preventing baserunners from scoring better than average? Other than the pitcher himself being intelligent in his plan when he gets guys on base, there is nothing to explain variations in M.O.B. performance. So unless yo're arguing that pitchers get dumber and smarter from year to year...single-season changes are meaningless.


There is a pretty strong correlation between guys having great K rates and great LOB% - that shouldn't be suprising. Outside of a few guys like Glavine, who have great LOB% without huge K rates (for the reasons Matt mentioned earlier), most all have very good DIPS stats. Many pitchers also seem to repeat good/bad LOB% numbers at a degree higher than BABIP and HR/FB at least. When a guy has a LOB% near 80% or under 65%, that is when I see "luck" written all over it. Most of the best pitchers consistanly go 72% - 77% year after year after year. Matt's "why" is the real question. I would bet most of it is a natural consequence of the TTO, with the rest being Glavine-like change of approach.

[brett]

There is a pretty strong correlation between guys having great K rates and great LOB% - that shouldn't be suprising. Outside of a few guys like Glavine, who have great LOB% without huge K rates (for the reasons Matt mentioned earlier), most all have very good DIPS stats. Many pitchers also seem to repeat good/bad LOB% numbers at a degree higher than BABIP and HR/FB at least. When a guy has a LOB% near 80% or under 65%, that is when I see "luck" written all over it. Most of the best pitchers consistanly go 72% - 77% year after year after year. Matt's "why" is the real question. I would bet most of it is a natural consequence of the TTO, with the rest being Glavine-like change of approach.

Ryan was wilder from the stretch and lost some of his K rate.

His leverage splits are not good either-best in low leverage and 104 in high leverage. I believe that leverage has nothing to do with "run scoring probability" but simply inning and score considerations. Seaver had a 94 high leverage split. I think that Ryan was mentally and physically affected with MOB.

And also by the way, whenever I try to estimate the effects of his 113 MOB split, I get it cutting his 123 DNRA+ down to the 109-112 range! I'd like someone to give me another stat based estimate.

If you get hit 13% harder (OPS+) with men on base-the only time someone can score except for a home run, you will give up ALMOST 13% more runs as I see it.

[Matthew C.]

And also by the way, whenever I try to estimate the effects of his 113 MOB split, I get it cutting his 123 DNRA+ down to the 109-112 range! I'd like someone to give me another stat based estimate.

.

But that would put him right back at his Baseball Reference ERA+ which doesn't account for what we know includes poor defense, poor pen and unjust park factors. If you accept his true performance to be 109-112, you are saying that his defense, park, and bullpen are neutral. That is why I think a half-way adjustment is probably pretty close to reality - 118 or so. His ERA+ is over penalizing a handful of points points due to the previously mentioned factors and DERA+ is crediting him a handful of points-too-many due to your contention. Big picture, a 117-118 guy with 5,200 IP is far below "greatest pitcher ever" that is is often cited by casual fans, but well above the "most overrated guy ever who was barely good enough to post records better than his mediocre teams", popular with members here.

[SavoyBG]

Umm...

Duh?

Of course great seasons typically come with good performance with men on base. The ERA tends be lower when you...um...don't give up hits with guys on base. Yep. Check.

That doesn't mean the year to year fluctuations in performance in that category are anything more than random. Think about this, Brett...when I approach a question like this, I start off by going to first principles...what physical explanations could there be for sustainable skill in preventing baserunners from scoring better than average?

Very simple, starting pitchers need to pace themselves and so they let up some when the face a batter with nobody on, especially when it's a batter who is not a huge HR/XBH threat. They need to conserve some energy for when they really need to bear down, with runners on base.

This was especially true in Ryan and Seaver's day, when starters were expected to complete most of their games. It may not be as true now, when startes are only expected to go 5 to 7 innings, and are being told to go as hard as they can for as long as they can with middle relievers stacked up to take over when the starter tires.

Pitchers were more likely to avoid walks and throw pitrches down the middle with nobody on and an out or two than they would with runners on base, Check it, I'm sure most pitchers have higher walk rates with runners on base.

The pitcher is not a Strat-O-Matic card that throws pitches randomly all game with results that are totally random from batter to batter.

They (attempt to) manage the game, based on situations.

[SavoyBG]

But that would put him right back at his Baseball Reference ERA+ which doesn't account for what we know includes poor defense, poor pen and unjust park factors. If you accept his true performance to be 109-112, you are saying that his defense, park, and bullpen are neutral. That is why I think a half-way adjustment is probably pretty close to reality - 118 or so. His ERA+ is over penalizing a handful of points points due to the previously mentioned factors and DERA+ is crediting him a handful of points-too-many due to your contention. Big picture, a 117-118 guy with 5,200 IP is far below "greatest pitcher ever" that is is often cited by casual fans, but well above the "most overrated guy ever who was barely good enough to post records better than his mediocre teams", popular with members here.

Ryan's longevity makes him great, but not near as great as many fans think. He's not in the frst tier of all time greats. He did not have a great peak. His greatness lies in being good to very good for a real long time, sort of the Pete Rose of pitching.

[SABR Matt]

Brett...the flaw in your logic is in using Ryan's self-relative OPS without also making that ratio LEAGUE relative. Because all pitchers (the group that contains all pitchers in any given league/year) have a split that favors the hitter with people on base. Check the league OPS split. It's usually about 6%. So Ryan isn't 13% worse than the average pitcher at preventing runs with people on base...(average pitching is what matters when computing DNRA+). You need the odds ratio method to compare Ryan's split to the league split and use that result to make a guess at how many runs that should cost him.

[SABR Matt]

Very simple, starting pitchers need to pace themselves and so they let up some when the face a batter with nobody on, especially when it's a batter who is not a huge HR/XBH threat. They need to conserve some energy for when they really need to bear down, with runners on base.

This was especially true in Ryan and Seaver's day, when starters were expected to complete most of their games. It may not be as true now, when startes are only expected to go 5 to 7 innings, and are being told to go as hard as they can for as long as they can with middle relievers stacked up to take over when the starter tires.

Pitchers were more likely to avoid walks and throw pitrches down the middle with nobody on and an out or two than they would with runners on base, Check it, I'm sure most pitchers have higher walk rates with runners on base.

The pitcher is not a Strat-O-Matic card that throws pitches randomly all game with results that are totally random from batter to batter.

They (attempt to) manage the game, based on situations.

Thank you once again for assuming I'm treating all players like strat-o cards and not seeing the bigger picture. In fact I've used your argument as a reason why ERA+ figures are higher for modern pitchers than for deadball stars. I think today's pitchers throw at max effort more often than they used to when they had to throw more innings. But all of your comments completely missed the point I was making.

What causes Ryan to go from being 10% better than himself with people on base one year...to 30% worse the next...and back again? What physical mechanism explains inter-annual variability in the pitcher's performance in clutch situations?

You won't find one...because it does not exist. Saying that does not deny the human element of the game...it simply recognizes that part of the reality of baseball is random variation in events around real skills that take longer than a single season to express themselves in the data itself.

I have no doubt that there are some pitchers who are better with men o base than others...and I've cited at least two possible explanations for it while STLCards offered a third (K rate limiting productive outs). That doesn't mean every run is absolutely a reflection of a pitcher's ability and not random chance. I think it is an important part of my "job" as a sabermetrician to try to see the pattern and ignore the noise to get to a greater understanding. I believe Ryan's ERA+ is badly misleading...in fact, I think Nolan Ryan and Mike Mussina are both great case studies for why using career ERA+ in isolation can lead to inaccuracies in assessment that, while not GIANT, are IMPORTANT.

[SavoyBG]

Thank you once again for assuming I'm treating all players like strat-o cards and not seeing the bigger picture. In fact I've used your argument as a reason why ERA+ figures are higher for modern pitchers than for deadball stars. I think today's pitchers throw at max effort more often than they used to when they had to throw more innings. But all of your comments completely missed the point I was making.

What causes Ryan to go from being 10% better than himself with people on base one year...to 30% worse the next...and back again?


Probably just a relatively small sample. He was likely only facing 300 to 400 batters with runners on base in his big seasons. You could also have a year where he's better simply because he had more 3B only and bases loaded situations where he pitched from the windup rather than the stretch. That would account for a physical reason why he was better in a particular year.

Pitchers will normally wind up with a runner on 3B only, and also with 2B and 3B or bases loaded, especially with two outs, where they figure if they get the batter out it doesn't matter much if the runners get a big jump.

[SABR Matt]

Thank you for proving my point FOR me. Makes my job a lot easier.

Yes indeed...the inter-annual fluctuations PROBABLY ARE due to small sample size noise. Meaning you cannot use ERA for season to season analysis of the real value contributed to total team defense by a pitcher because a significant element is generated by the luck of the draw on such small samples. Now over the course of 10+ years, you will start to get an idea of who was better than normal or worse than normal with MOB, but that should be left to objective analysis to try to tease apart how much of a player's performance was altered by contexts outside of his control and how much of it was legitimately generated by his ability. It is NOT a good assumption to claim that ERA+ is definitely accurate after enough innings are thrown.

[SavoyBG]

Thank you for proving my point FOR me. Makes my job a lot easier.

Yes indeed...the inter-annual fluctuations PROBABLY ARE due to small sample size noise. Meaning you cannot use ERA for season to season analysis of the real value contributed to total team defense by a pitcher because a significant element is generated by the luck of the draw on such small samples. Now over the course of 10+ years, you will start to get an idea of who was better than normal or worse than normal with MOB, but that should be left to objective analysis to try to tease apart how much of a player's performance was altered by contexts outside of his control and how much of it was legitimately generated by his ability. It is NOT a good assumption to claim that ERA+ is definitely accurate after enough innings are thrown.

But had you thought of the fact that pitchers do pitch from the windup with men on base, depending upon what bases the runners are on and how many outs there are?

Does the play by play info indicate whether or not a pitcher is winding up on each pitch?

[SABR Matt]

No...but the PBP data is broken down by starting base/out state and you can generally predict which motion the pitcher will be using based on that info alone except in some rare cases.

That's probably the split I'll prefer using (rather than just men on base)...wind-up situations vs. stretch situations, so that's a good point.

[SavoyBG]

No...but the PBP data is broken down by starting base/out state and you can generally predict which motion the pitcher will be using based on that info alone except in some rare cases.

That's probably the split I'll prefer using (rather than just men on base)...wind-up situations vs. stretch situations, so that's a good point.


I don't think it's just rare cases where it's hard to know if the pitcher is in the stretch or not. This is something that you notice when you are umpiring because it has a big effffect on what a pitcher can or cannot do in looking for a balk.

There are also many relief pitchers who NEVER pitch from a windup, even with nobody on.

I would say that it's not much more than 65/35 as to whether or not a pitcher will wind up or stretch with runners on 3B, 2B & 3B, or bases loaded, even with two outs. The score will be a factor in this too. A team with a 2 or 3 run lead is more likely to have a pitcher in the stretch with two outs and bases loaded. It also likely varies from pitcher to pitcher as many times the manager will tell the pitcher that it is hhis option whether or not to stretch.

Many pitchers feel that they have better control from the stretch as there is less activity with their motion, less chance for something to go wrong. Better control, but less velocity and movement.

[SABR Matt]

Even if you're correct (and you very well could be)...you can still look at the base/out state splits most likely to result in changes in the pitching motion. For guys who don't go to the stretch, you shouldn't see significant differences in their results...and that's just as important to know as whether they actually pitched from the stretch or not...or as whether pitching from the stretch affected some other guy's performance.

I think it's pretty clear that Ryan's command got slightly worse (resulting in a slight uptick in walk rate and HR rate) pitching from the stretch. I think it's pretty clear that Glavine was completely unphased by the stretch and that his pitching intelligence led him to find ways to maximize his chances of holding runners (same for Andy Pettitte). And you'll see all of this by looking at the DEFENSE INDEPENDENT statistics in base/out situations most conducive to changes in approach or mechanics.

[brett]

Another factor with Ryan was his poor SB allowed%. I think it was in the upper 70s%. Some suggested to him that he pitch from the windup all the time because opponents couldn't steal much more against him anyway.

And the CS rate surely could affect the defenses and it would be due to Ryan.

[Matthew C.]

Another factor with Ryan was his poor SB allowed%. I think it was in the upper 70s%. Some suggested to him that he pitch from the windup all the time because opponents couldn't steal much more against him anyway.

And the CS rate surely could affect the defenses and it would be due to Ryan.

Yes - he was definitely horrible at this aspect - but both ERA+ and DERA+ already account for this.

[SavoyBG]

SavoyBG just did a spectacular job proving himself ill informed on this issue, BTW. Not to belabor this point, but Pettitte was clearly outpitching Saunders from the DIPS perspective...the difference was obvious by the end of the second inning. The fact that the score went to 1-0 by the end of the third didn't phase me...I knew Saunders was pitching a good notch worse and the Yankees did indeed eventually get to Saunders while Pettitte was busily pitching a gem.

The sad thing is...I don't think Savoy will ever see why I could tell the difference. I doubt he's ever going to understand the value of missing bats without walking guys as an indicator of pitching skill that translates to sustained success and real value.

LOL...you're delusional if you think that you spot more things from watching a game than I do. My post was not sarcastic, I was agreeing that in this case despite not allowing a run yet it was clear that Saunders was not pitching well.

[brett]

That's not how I read this. I read in this that you mean that someone mindlessly looking at DIPS might take him out, but a real manager is going to look past that.


Saunders hasn't allowed a run yet after three innings, but it's time to get him out of there. With 6 baserunners already Scoscia just checked his PCA rating for the game and it's not very good. With a WHIP of 2.00 he'as actually pitching very badly and should be replaced now.....right Matt?

[brett]

Yes - he was definitely horrible at this aspect - but both ERA+ and DERA+ already account for this.

For the steals but not for the effect on defensive positioning right?

[brett]

Even if you're correct (and you very well could be)...you can still look at the base/out state splits most likely to result in changes in the pitching motion. For guys who don't go to the stretch, you shouldn't see significant differences in their results...and that's just as important to know as whether they actually pitched from the stretch or not...or as whether pitching from the stretch affected some other guy's performance.

I think it's pretty clear that Ryan's command got slightly worse (resulting in a slight uptick in walk rate and HR rate) pitching from the stretch. I think it's pretty clear that Glavine was completely unphased by the stretch and that his pitching intelligence led him to find ways to maximize his chances of holding runners (same for Andy Pettitte). And you'll see all of this by looking at the DEFENSE INDEPENDENT statistics in base/out situations most conducive to changes in approach or mechanics.

Why was Maddux so bad in the MOB split?

[brett]

Thank you for proving my point FOR me. Makes my job a lot easier.

Yes indeed...the inter-annual fluctuations PROBABLY ARE due to small sample size noise. Meaning you cannot use ERA for season to season analysis of the real value contributed to total team defense by a pitcher because a significant element is generated by the luck of the draw on such small samples. Now over the course of 10+ years, you will start to get an idea of who was better than normal or worse than normal with MOB, but that should be left to objective analysis to try to tease apart how much of a player's performance was altered by contexts outside of his control and how much of it was legitimately generated by his ability. It is NOT a good assumption to claim that ERA+ is definitely accurate after enough innings are thrown.


OK but sample size can't be the problem with MOB split fluctuations looking at DIPS or OPS+ for example. Well it does explain most of the fluctuations, but pitchers have as many BF with MOB as with the bases empty.

And perhaps the MOB is hurting the team defense, and can be seen in how it affects other pitchers as well.

[brett]

But that would put him right back at his Baseball Reference ERA+ which doesn't account for what we know includes poor defense, poor pen and unjust park factors. If you accept his true performance to be 109-112, you are saying that his defense, park, and bullpen are neutral. That is why I think a half-way adjustment is probably pretty close to reality - 118 or so. His ERA+ is over penalizing a handful of points points due to the previously mentioned factors and DERA+ is crediting him a handful of points-too-many due to your contention. Big picture, a 117-118 guy with 5,200 IP is far below "greatest pitcher ever" that is is often cited by casual fans, but well above the "most overrated guy ever who was barely good enough to post records better than his mediocre teams", popular with members here.

The numbers are a little confounding. Ryan actually produced a 111 ERA+.

We know that defenses, parks, pen etc HURT his ERA+. He did ACTUALLY PITCH better than 111 because he really produced 111 and had those disadvantages.

But his MOB split is poor. Nevertheless to bring him back down to 111 with defense/park adjustments would mean that he should have been worse than 111 in the first place!

Why can't we just translate an AVERAGE pitcher to HIS setting rather than the other way around. If an average pitcher would have had a 91 ERA+ in his setting and he had 111 then he would have had a 122 ERA+ ie.

[SABR Matt]

That's not how I read this. I read in this that you mean that someone mindlessly looking at DIPS might take him out, but a real manager is going to look past that.

Yep...I read the same. Nice try by Savoy to get off the hook for his blunder though.

[SABR Matt]

Why was Maddux so bad in the MOB split?

He was a contact pitcher...which means even his trickery was bound to produce more lucky singles with guys on base than your average pitcher.

Also...a big chuck of his increased OPS with guys on base is coming from an increased walk rate. And you can see that most of the increased walk rate is happening with RISP...not just men on...which means a lot of those walks were probably either intentional or semi-intentional. He was famously a strategic pitcher in the same way that Glavine was.

[SavoyBG]

That's not how I read this. I read in this that you mean that someone mindlessly looking at DIPS might take him out, but a real manager is going to look past that.

See my followup post.

You didn't need stats to see that Saunders was not pitching well and had just gotten lucky through three.

[SABR Matt]

For the steals but not for the effect on defensive positioning right?

Possibly a good point...but I suspect that impact is going to be minor. The defense is, if anything going to play a less distorted set of positions once the runner has successfully stolen second than when he's at first...so...any losses Ryan suffered due to the first baseman having to play close to the bag more often might be cancelled by gains in defensive positioning after he already allowed that runner to move up.

[leewileyfan]

What causes Ryan to go from being 10% better than himself with people on base one year...to 30% worse the next...and back again? What physical mechanism explains inter-annual variability in the pitcher's performance in clutch situations?

You won't find one...because it does not exist.

The sabermetric dismissal of many critically nuanced elements of MLB performance as "randomness," "noise," and "luck" is astounding.

In a "science" so dedicated to microscopic investigation of numbers, the human element, often glaring in the numbers, is often tossed in with the "noise," a recognizable nuisance in the regression process.

Case in Point: Nolan Ryan:

Career, Ryan was most dominant in the following situations:

Base State...........K/BB Ratio.......BA/OB%/SLG%..............TB/H

- - - 2.15................191/.298/.279.............1.462

- X - 1.84................184/.332/.272.............1.479

- X X 1.43................210/.356/.263.............1.253

X X X 3.11................207/.274/.318..............1.540

- - X [2 Out] 2.07.................200/.307/.288.............1.096

1. For a pitcher with control issues K/BB ratios @ 2 or more may be astounding; but that is not the key element. except at extremes > 2-1. However, a runner at first base is a real distraction; and even though his K/BB rate jumps to 2.4, the .240/.317/.356 numbers betray the distraction, along with the rise to 1.483 bases per hit. It is as if Ryan grudgingly cedes second base to the base stealers to rid himself of the distraction.

2. With a runner on 2B, Ryan is in one of his top three performance "zones." His K/BB drops to 1.51, due in large part to the option of putting a man on first to set up the force play all-around. His hitter line drops to .184/.332/.272.

3. In a - X X situation, Ryan's K/BB ration drops to his career situational low @ 1.43, again in light of the IBB option. However, the batting line = .210/.356/.263; and his TB/H drops to 1.253 bases @ 14.2% below his career average @ 1.460. Again, it seems Ryan recognizes his BB nemesis as his ally; and it ceratainly appears he bears down in this situation.

4. X X X reveals the full bear-down Ryan both near his best [3.11-1 K/BB ratio] and .207/.274/.318 line. He knows he has no place to put the batter and falls back on his K. However, the human element shows through: His TB/H = 1.540, highest of all situational numbers @ 1.540 bases per hit. Rather than walk in a run, he's grooving it.

5. The psychology of base-out situations with Ryan is best studied in the
- - X situation[s]:

@ - - X < 2 Outs = 1.69 K/BB; .258/.337/.367; 1.423 TB/H

@ - - X 2 Outs = 2.07 K/BB; .200/.307/.288; 2.07 K/BB; 1.096 TB/H [lowest situational, career

................. looks like two different guys.

This stuff is random, noise, luck? Don't think so.

1. With none on, Ryan has his full windup and is unthreatened in any way.

Saying that does not deny the human element of the game...it simply recognizes that part of the reality of baseball is random variation in events around real skills that take longer than a single season to express themselves in the data itself.

[SABR Matt]

Of course leewiley...if you had any reading comprehension skills, you would know that I was never arguing that the CAREER TOTALS in the situational splits were random. That's what you're giving me...career totals. I *AGREE!!* that a career as big as Ryan's will show patterns that can be understood and explored...that's a very good thing to be studying. What I was saying was that SEASON TO SEASON variation in metrics like the ones you just showed were noise...the runner at first base situation is going to be distracting to Ryan just the same when he's 35 as when he's 34 or 36. There's no rational or logical explanation for why Ryan's split with a runner at first base fluctuated about his career norms from year to year...that's just random variation.

[leewileyfan]

Of course leewiley...if you had any reading comprehension skills, you would know that I was never arguing that the CAREER TOTALS in the situational splits were random.

I was going to preface my post with the opener "In the sincere hope that no predicatble personal attack will repond," OR "At the risk of off-topic personal attack," ..... but I dumped that, foolishly hoping for better.

The Emperor of this board. SABR MATT], eternally bound to address persons before topics, has just modeled his new wardrobe.

Since you are always handing out your resume, let me tell you about me and reading comprehension:

-For 10+ years, I taught English on the secondary level, earning a chairmanship in 1967;

-Much of my graduate level study had to do with reading problems, interpretations, eye exercises and effects on comprehension, remedial practices, and enhanced speed reading;

-I had one year of experience in customizing reading programs for a private sector firm specializing in reading methods, enhancements, comprehension and effective remedial techniques;

You have repeatedly dismissed pitching nuances as noise, citing a single season or variations among several seasons as random noise or luck, lacking an identifiable, representational, definable core, at one time seeing 10 years or career supplying enough data to make a case for pitcher values, as defined by certain sabermetric measures. At the same time, you spout the elegance of sabermetric predictive values.

You can't seem to comprehend your own message, preferring instead to respond to posters with "duh?" reading challenges, or sarcastic "got it" snippets.

The more you exchange with others, the more you reveal your emptiness.


That's what you're giving me...career totals. I *AGREE!!* that a career as big as Ryan's will show patterns that can be understood and explored...that's a very good thing to be studying. What I was saying was that SEASON TO SEASON variation in metrics like the ones you just showed were noise...the runner at first base situation is going to be distracting to Ryan just the same when he's 35 as when he's 34 or 36. There's no rational or logical explanation for why Ryan's split with a runner at first base fluctuated about his career norms from year to year...that's just random variation.

That is Nolan Ryan according to SABR Matt. Others, more inclined to exploration of situational pitching from the perspective of overall career, are more open to spotting seasonal-age trends.

What is random is the volume of occurrence rates of situations among seasons and how those situations came into being:

X X - hit batsman + walk

X X - walk + booted DP ball

X X - dropped fly for 2B error + IBB

X X -missed 3rd strike + single, runner stopping at second

X X -single + steal + BB

Within each situation, season to season, there are min-dramas that create the situations you find either numerically valuable or the junk of ransom noise.

Comprhension essentially lies in seeing the forest for the trees.

[Matthew C.]

The numbers are a little confounding. Ryan actually produced a 111 ERA+.

We know that defenses, parks, pen etc HURT his ERA+. He did ACTUALLY PITCH better than 111 because he really produced 111 and had those disadvantages.

But his MOB split is poor. Nevertheless to bring him back down to 111 with defense/park adjustments would mean that he should have been worse than 111 in the first place!

Why can't we just translate an AVERAGE pitcher to HIS setting rather than the other way around. If an average pitcher would have had a 91 ERA+ in his setting and he had 111 then he would have had a 122 ERA+ ie.

I am raising his "ERA+" 7-8 points due to park/defense/pen. I don't need to do anything with MOB/SB% since it is already captured in ERA+ (hence his ERA+ being much lower than his OPS+A would indicate) since ERA+ deals with real earned runs, not MOB/running game neutralized runs. Only if you started with OPS+ would you need to increase for park/defense/etc. and penalize for MOB, etc.

[Matthew C.]

Here is something I looked at about 3 years ago - might shed a little light on the walking with MOB issue...

July 2006: The notion of a pitcher giving up a larger % of his walks with men on base is intreging to me.

I looked at every Hall of Fame pitcher that had available splits as well as 15 guys not eligable or not elected (John, Kaat, Blyleven, Morris,Tiant, Stieb, Cone, Maddux, Clemens, Glavine, Johnson, Martinez, K.Brown, Schilling, and Mussina) to see how many of them had a "don't give in and don't allow XBhit" mentality. Here is what I found...

There were only 7 guys (out of near 40) who gave up more walks with runners on base than without (and their % more of BB's w/RO that with nobody on)...

Robin Roberts 28% :eek:
Tom Glavine 25% :eek:
Curt Schilling 15%
Greg Maddux 14%
Tommy John 6%
Jim Kaat 6%
Warren Spahn 2%

Schilling and Maddux's K:BB ratios are even more incredible when you consider how many of many of their walks were "intentional" or "pitching around" situations with men on.

[SABR Matt]

Comprhension essentially lies in seeing the forest for the trees.

Yes...yes it does. What am I supposed to think of your comprehension when you argue from career totals in direct agreement with everything I've said in this and other threads about situational pitching analysis and claimed I was arguing something else?

I don't care what your resume is...and I have never handed out my resume here. I judge based only on the comments made at this forum...and you've shown a consistent refusal to accurately portray my position in debate. That's all I need to know.

As for Ryan, I don't see smooth age trends in his MOB vs. bases empty split. I see random ditty-bopping. It is certainly possible for a pitcher to learn how to cope with men on base as he matures...there's no sign of that with Ryan.

[SABR Matt]

Here is something I looked at about 3 years ago - might shed a little light on the walking with MOB issue...

July 2006: The notion of a pitcher giving up a larger % of his walks with men on base is intreging to me.

I looked at every Hall of Fame pitcher that had available splits as well as 15 guys not eligable or not elected (John, Kaat, Blyleven, Morris,Tiant, Stieb, Cone, Maddux, Clemens, Glavine, Johnson, Martinez, K.Brown, Schilling, and Mussina) to see how many of them had a "don't give in and don't allow XBhit" mentality. Here is what I found...

There were only 7 guys (out of near 40) who gave up more walks with runners on base than without (and their % more of BB's w/RO that with nobody on)...

Robin Roberts 28% :eek:
Tom Glavine 25% :eek:
Curt Schilling 15%
Greg Maddux 14%
Tommy John 6%
Jim Kaat 6%
Warren Spahn 2%

Schilling and Maddux's K:BB ratios are even more incredible when you consider how many of many of their walks were "intentional" or "pitching around" situations with men on.

Robin Roberts may be a pitcher who is underestimated by basic DIPS methods just as Glavine is. I say that because he always ran huge HR rates and that just KILLS you in the DIPS analysis...but a lot of his HRs were solo because he was pitching an order of magnitude more carefully with people on base. Which is good information to know.

[leewileyfan]

Yes...yes it does. What am I supposed to think of your comprehension when you argue from career totals in direct agreement with everything I've said in this and other threads about situational pitching analysis and claimed I was arguing something else?

VAPID and non-responsive. Errrrr, we are talking pitching, situational pitching, various "evaluators" and weightings and an entire range of observations, conclusions, and counter-conclusions. You contribute rhetorical questions, personal attacks and feigned innocence of context.


I don't care what your resume is...and I have never handed out my resume here. I judge based only on the comments made at this forum...and you've shown a consistent refusal to accurately portray my position in debate. That's all I need to know.

You came into my wheelhouse and I chewed on your armor. Simple as that. Your debating viewpoints are chameleon-like, arguing to the sentence in the thread in the moment. Your conclusions are always provided with a musical coda escape hatch; and you snipe at others. Period.


As for Ryan, I don't see smooth age trends in his MOB vs. bases empty split. I see random ditty-bopping. It is certainly possible for a pitcher to learn how to cope with men on base as he matures...there's no sign of that with Ryan.

Please provide us with your detailed situational splits on Ryan, 1966-1992 [or at least you immediate and reviewable source,

That shouls be a simple enough refernce to provide.

[SABR Matt]

I have the PBP database downloaded from retrosheet.org, for one...and I have baseball-reference.com as a quick at-a-glance tool for situational analysis.

You can think whatever you like about your debating technique. You certainly aren't afraid to blow hot air about how amazing you are. What I am waiting for is for you to provide some rational explanation as to why I should take year to year fluctuations in Ryan's situational splits seriously as real skill on his part and not random variation.

[leewileyfan]

I have the PBP database downloaded from retrosheet.org, for one...and I have baseball-reference.com as a quick at-a-glance tool for situational analysis.

You have Baseball-Reference from 1966-1992? EDIT~ Oh, yes, I'd perused that and found two notable trends for Ryan in years when he worked lots of innings. It's in there; but year-by-year really isn't all that necessary ... just attentive perusal.


What I am waiting for is for you to provide some rational explanation as to why I should take year to year fluctuations in Ryan's situational splits seriously as real skill on his part and not random variation.

1. Don't wait on my account. Put as simply as possible, looking at any pitcher's situational splits, year-by-year, will turn out to be a wearying, futile, and mind-numbing exercise if you enter such a task convinced in advance that it's all "noise."

2. An efficient researcher wouldn't have to do anything nearly so voluminous, if he took on the task with meaningful situations being the focus ... the rest as probable noise.

3. An efficient researcher would always have sets of pitchers of various strengths, weaknesses and generations, having already noted key elements from prior researches. This would reveal trends by handedness, style [finesse or power], and readily available references applicable to all decades of play covered.

Then the data reveals itself, not some self-fulfilling bias carried into the effort.

[brett]

How can the MOB split be noise when it accounts for half of a pitchers BFs? If that's the case, then aren't all season to season variations noise?

[SABR Matt]

A fair question, Brett. But again...I go back to first principles...give me a reason why Nolan Ryan (or anyone else) could pitch with exactly the same overall skills and similar results in the defense independent analyses...and yet legitimately cause his MOB split to wobble up and down like a yo-yo due to some direct actions he took and I'll be more than happy to consider it.

On some level, yes, MOST year to year fluctuations in performance are luck-driven (if they aren't related to aging or injury)...but this particular split is independent of overall competence in the same way that batters' BABIP (relative to their career norms) is independent of overall competence. I suppose you could argue on some level that those fluctuations (like the fluctuations in BABIP for hitters) are real accomplishments and should be recorded as such. If you're interested in just recording precisely what things a pitcher did to contribute to his team's cause while still weeding out biases like defense and park...you could alter DNRA by computing it with respect to each base/out state...a situational DNRA would not be as predictively valuable as the current model but it would have the advantage of showing exactly what the pitcher did that his fielders and park did not cause.

And I would still expect there to be significant divergences from ERA+ in the final analysis of the situational version of DNRA, BTW. For reasons I've already elaborated on.

[leewileyfan]

However, pitching is an element of defense and vice versa; so that methodologies that isolate the two may be counterproductive in determining relative values of the two forces.

Under that scenario, much discarded as junk is potential Shinola in an approach that sees the whole, then the parts.

[Matthew C.]

From a practical standpoint, here is how I can see somebody using DIPS in-game (since the question was brought up in a round-about-way earlier):

Lets take two pitchers, we will call them Chris Carpenter and Todd Wellemeyer.

In 2 innings of work, Wellemeyer goes:

I1: FO, 1B, BB, FO, BB, LO
I2: FO, BB, 1B, GO, FO

In 2 innings of work, Carpenter goes:

I1: SO, IF1B, 2B (ER), SO, PO
I2: SO, SO, GO

Through 2 innings, Wellemeyer has given up 5 runners and hasn't gotten a single strikeout. He is wild, and only a linedrive right at a fielder is the difference between 0 ER allowed and 2 ER allowed. He keeps giving up flyballs, and if that continues, it is only a mater of time before one leaves the park. Through 2 innings, Carpenter has carved up 4 batters with ks, gotten weak contact on a popout, and hasn't walked a single batter. He gave up an infield hit and one solid double.

So, can we say that Wellemeyer up to that point has pitched better?
Even if we can, breaking down strike/ball ratio, etc. can we expect Carpenter to continue giving up more runs than Wellemeyer? DIPS would say, no, Carpenter is more likely to continue pitching well, while Wellemeyer is likey to fall apart. Wouldn't we all expect Carpenter to continue pitching well, while Wellemeyer is "due" for all of those baserunners to come back and kill him in any inning? Of course we would. The fact that he is only getting contact means that sooner or later, a few of those balls will likely find a whole and all of those walks will come in to score. That is why there is a minimal standard of K/9 that pitchers have to have to be moderately successful.

[SABR Matt]

Or...a more directly strategic example (something a manager might actually have to make a decision on)...

You've got King Felix on the hill and he's gone 7 innings allowing 8 hits (of which 5 were on the ground and none were struck particularly hard), 2 runs, 1 walk and 9 Ks. He's thrown 98 pitches in seven and he's come out to start the 8th inning. In the bullpen today, you don't have Mark Lowe available because he just pitched 1.1 innings and needed 36 pitches yesterday, so you've got Sean White and Miguel Batista up and throwing.

Now the Mariners are leading 3-2 and Felix starts the 8th with an infield single (more weak contact), a brutal punch out on his yellow hammer and grounder down the line for a double. Runners at second and third, one out...he's up to 112 pitches now but still throwing hard. You call down to Wetteland and he says Sean White and Miguel Batista are both ready to go if you need him.

White's DIPS line:

42 IP, 19 K, 18 BB, 3 HR (2.85 ERA)

Batista's:

54 IP, 40 K, 33 BB, 1 HR (4.10 ERA)

Both guys are groundball pitchers more than flyball pitchers...neither guy is overwhelming with his K/BB, but Miguel is more likely to get a strikeout if you need it (and also more likely to walk people).

There is a third option...let Felix try to pitch his way out of trouble, since he's still got good stuff and has been striking guys out left and right.

You've got Curtis Granderson, Placido Polanco, and Miguel Cabrera due up. What do you do?

A typical fan would be yelling at the TV "he's tired...get your set-up guy in there!!" They'd be calling for Sean White. I believe, however, that the correct decision when you need to hold a 1-run lead is to EXCLUDE White (because he's not likely to hold that runner at third)...you have the option to walk Granderson (if Felix doesn't think he can strike him out...and he is prone to the K, so maybe you go with plan A and pitch aggressive to him) and take your chances with Polanco, trying to get a groundball hitter to hit one on the ground for a double play. If he thinks he can get the punchy or the DP grounder and has enough left in the tank...stick with Felix...if he gives you any indication he's toast...go with Batista - as scary as that proposition might be given that it flies in the face of traditional bullpen management (White has a good ERA this year...he's been your 7th inning guy most of the season) and/or pitch count management for the starter.

Notice, as well...my decision is not limited to strat-o style statistical comparisons...I asked Felix how he was...I trust the coaching staff to know when he's legitimately done...and I considered groundball/flyball tendencies as well, which are also an attribute that is defense independent and critical to this kind of decision.

[brett]

From a practical standpoint, here is how I can see somebody using DIPS in-game (since the question was brought up in a round-about-way earlier):

Lets take two pitchers, we will call them Chris Carpenter and Todd Wellemeyer.

In 2 innings of work, Wellemeyer goes:

I1: FO, 1B, BB, FO, BB, LO
I2: FO, BB, 1B, GO, FO

In 2 innings of work, Carpenter goes:

I1: SO, IF1B, 2B (ER), SO, PO
I2: SO, SO, GO

Through 2 innings, Wellemeyer has given up 5 runners and hasn't gotten a single strikeout. He is wild, and only a linedrive right at a fielder is the difference between 0 ER allowed and 2 ER allowed. He keeps giving up flyballs, and if that continues, it is only a mater of time before one leaves the park. Through 2 innings, Carpenter has carved up 4 batters with ks, gotten weak contact on a popout, and hasn't walked a single batter. He gave up an infield hit and one solid double.

So, can we say that Wellemeyer up to that point has pitched better?
Even if we can, breaking down strike/ball ratio, etc. can we expect Carpenter to continue giving up more runs than Wellemeyer? DIPS would say, no, Carpenter is more likely to continue pitching well, while Wellemeyer is likey to fall apart. Wouldn't we all expect Carpenter to continue pitching well, while Wellemeyer is "due" for all of those baserunners to come back and kill him in any inning? Of course we would. The fact that he is only getting contact means that sooner or later, a few of those balls will likely find a whole and all of those walks will come in to score. That is why there is a minimal standard of K/9 that pitchers have to have to be moderately successful.

As a quick aside here, what are the worst 15+ K games by a pitcher in history? I don't expect there are many poor ones, but there are times I worry when a pitcher only seems to be getting guys out with the K. Maybe its because he is using a lot of pitches. When you get a pitcher in the first who gives up three quick line-drive outs on 5 pitches, the announcers usually talk about how efficient he was, but he actually was lucky.

How many zero K shutouts have been thrown in the last 50 years?

(Oddly, I seem to remember Clemens throwing a 1 K shutout in '86, the year of his first 20 K game.

[SABR Matt]

I remember Ryan Franklin flirting with a no-no for a long while as a Mariner and going 8+ innings with only a few hits and 0 runs while fanning just 1.

[leewileyfan]

Or...a more directly strategic example (something a manager might actually have to make a decision on)...

You've got King Felix on the hill and he's gone 7 innings allowing 8 hits (of which 5 were on the ground and none were struck particularly hard), 2 runs, 1 walk and 9 Ks. He's thrown 98 pitches in seven and he's come out to start the 8th inning. In the bullpen today, you don't have Mark Lowe available because he just pitched 1.1 innings and needed 36 pitches yesterday, so you've got Sean White and Miguel Batista up and throwing.

Now the Mariners are leading 3-2 and Felix starts the 8th with an infield single (more weak contact), a brutal punch out on his yellow hammer and grounder down the line for a double. Runners at second and third, one out...he's up to 112 pitches now but still throwing hard. You call down to Wetteland and he says Sean White and Miguel Batista are both ready to go if you need him.

White's DIPS line:

42 IP, 19 K, 18 BB, 3 HR (2.85 ERA)

Batista's:

54 IP, 40 K, 33 BB, 1 HR (4.10 ERA)

Both guys are groundball pitchers more than flyball pitchers...neither guy is overwhelming with his K/BB, but Miguel is more likely to get a strikeout if you need it (and also more likely to walk people).

There is a third option...let Felix try to pitch his way out of trouble, since he's still got good stuff and has been striking guys out left and right.

You've got Curtis Granderson, Placido Polanco, and Miguel Cabrera due up. What do you do?

Matt, the difference between us is that I read this post through the quoted portion, since it posed a credible situation and provided numbers a manager would have as he considered a decision. I would gladly have participated in the playing out of your scenario [as I'm sure would several others her] and waited for the discussion to follow.

Then, you spoil it with this:


A typical fan would be yelling at the TV "he's tired...get your set-up guy in there!!" They'd be calling for Sean White. I believe, however, that the correct decision when you need to hold a 1-run lead is to EXCLUDE White (because he's not likely to hold that runner at third)...you have the option to walk Granderson (if Felix doesn't think he can strike him out...and he is prone to the K, so maybe you go with plan A and pitch aggressive to him) and take your chances with Polanco, trying to get a groundball hitter to hit one on the ground for a double play. If he thinks he can get the punchy or the DP grounder and has enough left in the tank...stick with Felix...if he gives you any indication he's toast...go with Batista - as scary as that proposition might be given that it flies in the face of traditional bullpen management (White has a good ERA this year...he's been your 7th inning guy most of the season) and/or pitch count management for the starter.

No discussion at all; with a condescending reference to the "typical fan," even though same had no opportunity to speak for himself.

FWIW, if I were the manager [and I hadn't yet made any trips to the mound]. I'd go to Felix and say, "Guy, you deserve a win here - give us just two more outs and we'll try to nail it down for you. This situation calls for "the man," they guy into the game rhythm and tossing a beauty. Just get us two more."

No way I go to either guy in the pen unless King tells me he's done. Then, I stall for an "unintentional" IBB, with Lowe up and throwing to get two outs - off season makes for long rests. If Lowe can't get ready & Felix is out of gas, I go with White.

[SABR Matt]

That line was not intended to be condescending. I've sat in enough sports bars to know what is commonly said in situations such as the one I posed. I certainly wasn't necessarily implying that the common fan is wrong all the time or stupid for his beliefs...I think having more data and more time spent studying the game does give a person more power to understand what he is seeing and make sounder judgments...that's just a fact of life in all areas.

As I said in my conclusion, I would stick with Felix if he told me he still had some left in the tank and my coaching staff agreed he was OK...I generally agree that this situation calls for the power arm in whom you trust. In the event that I can't get Felix through the 8th, I go with Batista over White for reasons stated in the original post - primarily, I think the three most likely outcomes with Batista (K, ground ball somewhere, BB) are less likely to be negative than the two most likely outcomes with White (ground ball somewhere, flyball somewhere). Lowe is not available...managers have to sometimes avoid going with their best arms today to make sure they have them tomorrow and the next day and my scenario reflected this.

I think viewing these kinds of situational decisions with an eye toward the DIPS paradigm can and does help a manager set up his relievers to have a higher chance for success and improves the results for his starters too. If you've got a starter in there who's given up 4 runs in the first on 6 consecutive hits, several of which were seeing-eye grounders or bloops, you don't panic and get the bullpen up...you let him try to work through it. Because he's pitching in some bad luck right now, and you want to give him some confidence and assure him that you are giving him every opportunity to improve the outcome of the day and keep working on it. If you've got the same result, but several of the six hits in a row were scalded, your reliever should have already been up and throwing after about the fourth hit and he's in the game now. Get him out of there before he makes it any worse...it's obviously not his day.

[Matthew C.]

lucky.

How many zero K shutouts have been thrown in the last 50 years?

.

56

Intersting (albeit not necessarily useful) info...

Only one in the 2000's - Scott Erikson 2002
Only three in the 1990's - Jim Bullinger 1997, Roger Bailey 1997, Zane Smith 1994

Only 5 from anybody even considered as a borderline HOF candidate:
Jim Kaat 1965, Tommy John 1973, Rick Reuschel 1973, Tommy John 1980, Roger Clemens 1987, Tom Glavine 1989
It is amazing to think that Clemens could go 9 innings and not strike anybody out - especially back in his flame-throwing days!

Repeats: Tommy John (2), Jerry Ruess (3), Ken Holtzman (3), Claude Osteen (2), Mike Caldwell (2), Al Fitzmorris (2)

Interestingly, of the 56 times it has been done, 25% of them have been done by multiple achievers. There seems to be a little more than just random fluctuation here. Being a low- K guy with a large subset of other run producing skills seems to be the non-luck part of the recipe. Not that any of this is surprising.

In 1965 Rich Beck had the most players reach base while throwing a SO with no K's, with 14 (9 hits and 5 walks). Neil Allen allowed the fewest with 2 hits in 1986. The only no-hitter was Ken Holtzman in 1969. Of the 56 games, only 7 had no walks.

[SavoyBG]

Ihe more you exchange with others, the more you reveal your emptiness.


Matt's the most knowledgable baseball man on the planet.....just ask him.

[SABR Matt]

I think it would be appropriate at this time to point out that while SavoyBG and leewiley have been relentlessly attacking me, I've been having a completely non-personal, completely rational, and very pleasant conversation with Brett, STLCards2 and occasionally other members will pop in with questions and comments.

Perhaps there's a reason for the difference in conversational tone, hm?

[SavoyBG]

I think it would be appropriate at this time to point out that while SavoyBG and leewiley have been relentlessly attacking me, I've been having a completely non-personal, completely rational, and very pleasant conversation with Brett, STLCards2 and occasionally other members will pop in with questions and comments.

Perhaps there's a reason for the difference in conversational tone, hm?

It's probably that you don't like when he and I keep proving your theories wrong.

[RuthMayBond]

It's been real.
It was fun.
It's no longer real fun.
I've had enough of the three of you, I think there's enough blame to go around.
Have a nice life, I'm probably not the only one who is unsubscribing to this thread :waving:waving:waving:waving:waving:waving:waving: waving

[brett]

56

Intersting (albeit not necessarily useful) info...

Only one in the 2000's - Scott Erikson 2002
Only three in the 1990's - Jim Bullinger 1997, Roger Bailey 1997, Zane Smith 1994

Only 5 from anybody even considered as a borderline HOF candidate:
Jim Kaat 1965, Tommy John 1973, Rick Reuschel 1973, Tommy John 1980, Roger Clemens 1987, Tom Glavine 1989
It is amazing to think that Clemens could go 9 innings and not strike anybody out - especially back in his flame-throwing days!

Repeats: Tommy John (2), Jerry Ruess (3), Ken Holtzman (3), Claude Osteen (2), Mike Caldwell (2), Al Fitzmorris (2)

Interestingly, of the 56 times it has been done, 25% of them have been done by multiple achievers. There seems to be a little more than just random fluctuation here. Being a low- K guy with a large subset of other run producing skills seems to be the non-luck part of the recipe. Not that any of this is surprising.

In 1965 Rich Beck had the most players reach base while throwing a SO with no K's, with 14 (9 hits and 5 walks). Neil Allen allowed the fewest with 2 hits in 1986. The only no-hitter was Ken Holtzman in 1969. Of the 56 games, only 7 had no walks.

I wonder with Clemens if his ability to get ahead in the count didn't still help.

Does anyone have BABIP rates with different ball/strike counts?

[leewileyfan]

I think it would be appropriate at this time to point out that while SavoyBG and leewiley have been relentlessly attacking me, I've been having a completely non-personal, completely rational, and very pleasant conversation with Brett, STLCards2 and occasionally other members will pop in with questions and comments.

Perhaps there's a reason for the difference in conversational tone, hm?

Enough is enough: the wounded "innocent" posing as the offended party.

Here, a bit like holding a mirror up to Medusa or publicizing Dorian's Gray's moldy portrait, is the tone exuded by SABR Matt on his fiefdom thread:

Dorian Gray’s Mirror


1. Indeed. And incidentally, it doesn't show much of an appreciation for the game to completely ignore what's happening now in your analysis work, tune in to a game you say you don't even pay attention to (to the point where you can't name many current players by your own admission) and then start leaping to unfounded conclusions as though anyone with half a brain couldn't possible disagree when in fact there are many people (who have entire brains! ) who have done some of the necessary early work involved in proving the merits of looking at team defense as a context who would indeed disagree with you.

If you don't like baseball enough to pay attention to it today...why are you here?

Wow ~ this courtesy because the poster clearrly stated that his focus was on MLB's historic context and players from earlier generations.

2.


Who in his/her right mind could possibly interpret those first two "grunts" as anything but a direct attack on the intellect of another poster's suggestion?

3. [quote]SavoyBG just did a spectacular job proving himself ill informed on this issue, BTW. Not to belabor this point, but Pettitte was clearly outpitching Saunders from the DIPS perspective...the difference was obvious by the end of the second inning. The fact that the score went to 1-0 by the end of the third didn't phase me...I knew Saunders was pitching a good notch worse and the Yankees did indeed eventually get to Saunders while Pettitte was busily pitching a gem.

By this stage of thread process, Matt had already belabored the topic.


The sad thing is...I don't think Savoy will ever see why I could tell the difference. I doubt he's ever going to understand the value of missing bats without walking guys as an indicator of pitching skill that translates to sustained success and real value.

On top of which ~ this zinger is added. Note: there is no reference to particular clues noted bt Matt that everybody else presumably missed.

4.
Yep...I read the same. Nice try by Savoy to get off the hook for his blunder though.

Belabored point ~ belabored yet again.

5.


This pip came my way. It got a very direct, clear response.

6. [quote]When was this book published? Because if it was after about 2002, the auther is lying in his claims about this being the first comprehensive catching analysis.

Hints at book not read.


And few sabermetricians would agree about Hartnett being the greatest of all time. I have him somewhere around 8th all time at the position...and that is based on a comprehensive analysis. :)

Choses to speak for "most" and offers his opinion ~ with pacifying smiley.

7.
leewiley...if you're [the author of the book being ripped - not I] going to declare yourself an expert and say you've made a breakthrough, then you'd better damned well do your research on what else has already been discussed by other sabermetricians. That's called being a GOOD SCIENTIST. If it's snarky to point out that this guy is about 8-10 years behind the forefront in sabermetrics and feel annoyed that he is nonetheless saying he's the first one to the top of the mountain...then I shall continue ot be snarky.

If you have ever been through the scientific peer review process then you'd know that people who claim to have made a breakthrough that has already been made five years ago and published for all to see get laughed out of academia and for GOOD REASON. Part of being a public figure, part of the responsibility you take on as a publisher is to say what is true and back it up with due diligence before you publish something for all to read. This clown didn't do his homework and I am well within my rights as a member of baseball's very young (admittedly) scientific community to be offended by him and to ridicule him for his lack of work.

Here, Matt & I agree completely. That was snarky ~ nasty even.

8.
I have not read the entire book...I don't need to. The summary of his methodology tells me it's a methodology that doesn't merit my time and energy. And his big sales pitch line "this is a breakthrough and no one has even attempted this before" pisses me off on a scientific level. Don't claim to be a scientist and then clearly fail to follow the scientific method. He skipped the SECOND STEP of good science...he didn't observe. He didn't do the research to study the problem/question he was trying to address. Slapped some crap together and then acted like he was a genius for doing it.

Book not read, crucified. Yet other books must be read in toto before critique is valid. I guess Matt decides which is which.




Noted. Again.

9. Let me spell this out very carefully so that your mind can process it and perhaps so that you can learn something. Nobody saying boo about the Japanese leagues doesn't mean nobody ANYWHERE...do you really imagine I haven't spoken to Jim Albright?

Ahhh, yes: speak slowly and enunciate, so that lesser minds can attempt to grasp the gravitas of what is being offered.

10. [quote]It's only misleading if you don't actually read the entire book. In the opening section of the first book, Shell defines what he means by the greatest hitter very specifically.

This is one of those in toto books, I guess.

Now, maybe we can return to brett's question about pitchers.

[brett]

[B]
2. [quoterUmm...

Duh?

Of course great seasons typically come with good performance with men on base. The ERA tends be lower when you...um...don't give up hits with guys on base. Yep. Check.

That doesn't mean the year to year fluctuations in performance in that category are anything more than random. Think about this, Brett..
[/QUOTE]

You have to carefully read context when trying to judge something like this. My question was effectively rhetorical and I took his response quite differently. "Duh" is not questioning my intellect. It implies that I am able to see the answer. What if someone asks a friend if they should go out with a Megan Fox and the friend responds "Duh"?

[leewileyfan]

You have to carefully read context when trying to judge something like this. My question was effectively rhetorical and I took his response quite differently. "Duh" is not questioning my intellect. It implies that I am able to see the answer. What if someone asks a friend if they should go out with a Megan Fox and the friend responds "Duh"?[/QUOTE]

Since you are/were the addressee, who am I to argue the point? However, the subsequent "errrrr" and the obvious reference to unquestionable facts would imply snark to readers not personally involved in the barbs exchanged.

Now, may we go back to your BB question?

[SABR Matt]

You have to carefully read context when trying to judge something like this. My question was effectively rhetorical and I took his response quite differently. "Duh" is not questioning my intellect. It implies that I am able to see the answer. What if someone asks a friend if they should go out with a Megan Fox and the friend responds "Duh"?

Correct Brett...I was using the "duh" in the friendly "thank you, captain obvious" sort of way, not the "you're an idiot" sort of way. This is why the internet is dangerous and generally not very hospitable. Because people make assumptions about the intentions of the other people to whom they are communicating (all in text) that would never happen if they could see the expression on everyone's face as they talked.

[SABR Matt]

The bottom line here is that I can tell when someone enters a conversation with a set agenda and a group of hard assumptions about me, and I don't like it. leewiley and Savoy are both, IMHO, always reading the most negative spin on any of my words because they believe there is zero inherent value in anything a "math geek" (to directly quote one of them) has to say about baseball and they think anyone who studies the sport scientifically is bound to completely ignore the human element of the game. So no matter what I say, they attack me and claim everything I say has the intent of calling them stupid.

If you enter a conversation with me the way STLCards2 and brett typically do - genuinely curious and looking for ways to improve upon what I've done...you get a COMPLETELY different kind of response from me. There's a reason. Most people don't like to be labeled and typecasted, and I'm no exception.

[leewileyfan]

The bottom line here is that I can tell when someone enters a conversation with a set agenda and a group of hard assumptions about me, and I don't like it.

Cite one instance where I opened a question or response to you, on a matter directly related to baseball and stats, [NOT defending myself or another being ground up in your sarcasm, ... with an agenda, a chip on my shoulder, a defensive tone or rude condescension. Just ONE.

I cited 10 for you without even trying.


If you enter a conversation with me the way STLCards2 and brett typically do - genuinely curious and looking for ways to improve upon what I've done...you get a COMPLETELY different kind of response from me.

I have NEVER attacked PCA or its findings, anywhere. What I have stated is that it is very difficult to locate, in any continuous presentation, anywhere via Google search or Yahoo search, even following links provided.
Last time I Googled, I gave up @ page 8.


You have identified two posters who are admittedly fans of PCA, generally trying to stay on topic and asking you direct questions for you to field. It's very hard to be extremely rude when you are responding to direct, polite questions.


There's a reason. Most people don't like to be labeled and typecasted, and I'm no exception.

We finally agree on something. You don't like it when you get responses in-kind to your own. That's how we feel, too. No one likes to be negatively typecast.

[Matthew C.]

Continuing the discusion on Ryan - I just saw a site which converted pitchers' Game Scores into W-L%, and it had Ryan's Game Score W-L record near 390-210. I never realized just how much Game Score overemphasized strikeouts.

[SABR Matt]

LOL

I'm not a fan of the game score, but I saw the SABR presentation where its secondary-proponent (a follower of James who created it) defended its use for posting corrected W-L records and he did have some good points about its surprisingly robust performance throughout the PBP era in correlating strongly with support/defense/park adjusted W-L.

[Matthew C.]

LOL

I'm not a fan of the game score, but I saw the SABR presentation where its secondary-proponent (a follower of James who created it) defended its use for posting corrected W-L records and he did have some good points about its surprisingly robust performance throughout the PBP era in correlating strongly with support/defense/park adjusted W-L.

Well, since it relies heavily on K and BB, it does hit on a few key DIPS components and it does neutralize for off. support, but when Ryan wins 380 games, Sutton wins 360 games, Mussina is 220-180, and when Andy Pettitte, Kenny Rogers and Jamie Moyer are below average pitchers (forget the HOF arguments), something is really wrong.

The site that I looked at did not adjust for era (which may explain why modern pitchers fair so poorly due to fewer IP) or park - so it was probably even worse that the presentation that you saw.

Oh yeah - Halliday is 210-190 or something!

[brett]

Did Mussina's defenses hurt him as much as his offenses helped his record?

I estimate Ryan would have had between 351 and 360 wins and 254 losses in a totally neutral setting.

[Matthew C.]

Did Mussina's defenses hurt him as much as his offenses helped his record?

I estimate Ryan would have had between 351 and 360 wins and 254 losses in a totally neutral setting.

I bet his W-L record is pretty close.

But we know for a fact that he is no .545 pitcher. And Ryan is no .650 pitcher. And Halliday is no .540 pitcher. And Wells, Moyer, and Pettitte are no .500 or below pitchers.

Game Scores double-counts K's, ignores park factors, pays zero mind to situational pitching, LOB%, etc., counts all hits as being equal (hurting guys who prevent XBhits), and does not account for era/IP accumulation. This explains why Bert Blyleven comes out looking better than Spahn, and why Sutton turns out looking better than Robin Roberts.

I bet there is a much better correlation between all-time Game Score ranking and finessness of pitchers than Game Scores and run- support. Almost all of the guys with the biggest GS drops compared to real W-L% are modern day, lower K guys. It has Chuck Finley, Kenny Rogers and Jimmy Key as average to below average too pitchers.

[SABR Matt]

Agreed on all counts STL. Pitching is definitely a much more complex task than the game scores make it, and Ryan is a good example of why.

Mussina did have a lot of offensive support, but he also is unfairly penalized by park factor analysis and by the Yankee defenses for which he pitched.

[brett]

I am interested in how PCA rates the following sets of pitchers:

First
Carlton
Perry
Niekro
Ryan

Second
Gibson
Koufax
Marichal
Whitey Ford
Jim Palmer

If you get some time I would like to see their year by year PCA scores. How would you rate the pitchers in each group? Is group 1 clearly better than group 2?

[jkberts1]

I am new to this forum, I think I hit a gold mine.

What exactly is PCA? Please give extensive details. Is there a database where I could download the data?

I am looking mainly for defensive ratings, at each position, from 1893 to present day.


Thank you for yuor time,


"Karl"

jkberts1@comcast.net

[SABR Matt]

Hello Karl...you may want to check your e-mail. :)

To get you started, I have sent along the awards data set, which gives top-tens for each league/year at each position for defense, and top 20s for batting wins and pitching wins (again for each league/year).

I'll send along the PCA Manifest is you're interested in the full description of the system...or some general commentary if you want a shorter explanation.

[SABR Matt]

I am interested in how PCA rates the following sets of pitchers:

First
Carlton
Perry
Niekro
Ryan

Second
Gibson
Koufax
Marichal
Whitey Ford
Jim Palmer

If you get some time I would like to see their year by year PCA scores. How would you rate the pitchers in each group? Is group 1 clearly better than group 2?

Do you want PCA wins from the original system or DNRA+ (the upgrade which includes more data from the PBP era and will be used in future PCA releases)?

[curveball]

Do you want PCA wins from the original system or DNRA+ (the upgrade which includes more data from the PBP era and will be used in future PCA releases)?

I believe Brett also wanted Blyleven thrown in there from the first group.

[brett]

Do you want PCA wins from the original system or DNRA+ (the upgrade which includes more data from the PBP era and will be used in future PCA releases)?

The second, as my main hope is to account for defense as accurately as possible. And I'd like Blyleven too, and your opinion in assessing them.

Did you mention that Neikro probably gets underrated in DNRA+?
(that's why I need an opinion too, and also looking at a players demonstrated level)

I noticed that Carlton doesn't look any better in career numbers than Perry, Niekro or Blylevin, but he almost always gets rated higher because of the big seasons. He is last or tied for last in ERA+, ERA+ for peak consecutive 3000 IPs and 2000 IPS.

IP
ERA+
ERA+ 3000 (best stretch of complete seasons totalling 3000+ IP)
ERA+ 2000 (same for 2000+ IP)


Carlton Niekro Perry Blyleven
IP 5217 1/3 5404 1/3 5350 1/3 4970
ERA+ 115 115 117 118
ERA+ 3000 124 125 133 127
ERA+ 2000 126 129 130 134


Carlton, Niekro, Perry, Ryan, Blylevin and Jenkins too if possible.

And Koufax, Gibson, Marichal, Palmer and Ford. I wanted Hubbell but I don't think the data is as accurate.

Thanks.

Palmer and Ford are high on a lot of people's list who don't like Smoltz, Schilling and Kevin Brown and I think that they both might actually be below those 3. Marichal might be too.

[brett]

2 quick questions about methodology Matt.

1) Do you take into account that a pitcher does not pitch equally against each team opponent and at home and away, or do you just have a single TEAM run environment for DNRA+

2) How do you calculate career DNRA+? Do you take the pitchers total career DNRA divided by his total environment or do you take the weighted average of each season's relative DNRA?

for a simple example here, lets say column 1 is the pitcher's DNRA and column 2 is the league and column 3 is the relative DNRA. Assume the same number of innings each year.

Season 1: 2 4 0.500
Season 2: 1 3 0.333
Season 3: 2 3 0.667
Season 4: 9 8 1.125
Season 5: 1 2 0.500

The average of column 1 is 3. The average of column 2 is 4.

The relative ERA would be 0.750 by taking the average of each column separately but:

The average of the relative ERA's is 0.625 which I believe to be more appropriate. Using method one to average, Season 4 gets weighted more than all the other seasons combined because the denomonator is 8.

[AstrosFan]

I'd rather read a page of PCA than watch an hour of ESPN, Law and Order reruns, just about anything.

I can tell my wife that I'm working on the computer. All she sees are names and stats. She thinks that PCA is the new state high school standards test and Clemente, Waner, and Kaline are some of my students:D.

I'm making popcorn tonight brotha!

(No pressure Matt if its not done tonight).

Every time brett makes a request, I think of this post.

[leewileyfan]

Well, since it relies heavily on K and BB, it does hit on a few key DIPS components and it does neutralize for off. support, but when Ryan wins 380 games, Sutton wins 360 games, Mussina is 220-180, and when Andy Pettitte, Kenny Rogers and Jamie Moyer are below average pitchers (forget the HOF arguments), something is really wrong.

I offer my effort purely on a FWIW basis. I have tried to normalize ERA over the period 1901-present so that all pitchers are allegedly projected against a level standard [4.50].

Sticking with pitchers emphasized here on recent posts, I've done a Pythag against career decisions to project career Wins vs actual career wins for each.

Pitcher..........Adj. ERA........Decisions.........Pythag %......E Wins.....Wins

Perry............. 3.82.............589............... .570.......... 336....... 314
Niekro............ 4.40.............592............... .512.......... 303........318
Ryan.............. 3.62.............616............... .607.......... 374........324
Blyleven.......... 3.85.............537............... .577.......... 310........287
K. Brown......... 3.13.............355............... .674.......... 239........211
Carlton........... 3.81.............573............... .582......... 334........329
Jenkins.............3.77..............510......... ...... .588.......... 300........284

Gibson..............3.51..............425......... ...... .622.......... 264........251
Koufax..............3.18..............252......... ...... .667...........168........165
Marichal............3.57..............385......... ...... .639...........246........243
E. Ford..............3.19..............342........... .... .666..........228.........236
Palmer...............3.50..............420........ ....... .624..........262........268

Hubbell..............2.42..............407........ ....... .776.........316........253

Pettitte.............3.66...............364....... ........ .602.........219.......229
Wells.................4.14..............396....... ......... .542.........214........239
Moyer...............3.96...............453........ ........ .564.........255........258
Finley................4.27..............373....... ......... .526........196........200

The biggest outliers would appear to be Ryan and Hubbell. However, when one factors in team offense and team defense behind each + [in Ryan's case] a control issue, the gaps lessen in full context:

-Hubbell's pitching staffs hovered @ slightly above average collectively for most of Hubbell's career. Subtracting Hubbell and recalculating the balance of the staff, pitching for the NYG was average to below-average, at best, without Hubbell.

-Defense behind Hubbell was below average for @ two-fifths of his career; average @ one-quarter; and above average for the balance.

-Batting runs behind Hubbell were @ 70 > LG average through most of his career. Collectively, we'relooking at a .500 team, 1928-1943 for the most part, without Hubbell. Re-inject Hubbell into the equations and we have a contender.

P.S. My line for Don Sutton is:

Adj. ERA [4.05]; Decisions [580]; Pythag Est. W% .552; Proj. W [320] Actual W [324]

[SABR Matt]

2 quick questions about methodology Matt.

1) Do you take into account that a pitcher does not pitch equally against each team opponent and at home and away, or do you just have a single TEAM run environment for DNRA+

2) How do you calculate career DNRA+? Do you take the pitchers total career DNRA divided by his total environment or do you take the weighted average of each season's relative DNRA?

for a simple example here, lets say column 1 is the pitcher's DNRA and column 2 is the league and column 3 is the relative DNRA. Assume the same number of innings each year.

Season 1: 2 4 0.500
Season 2: 1 3 0.333
Season 3: 2 3 0.667
Season 4: 9 8 1.125
Season 5: 1 2 0.500

The average of column 1 is 3. The average of column 2 is 4.

The relative ERA would be 0.750 by taking the average of each column separately but:

The average of the relative ERA's is 0.625 which I believe to be more appropriate. Using method one to average, Season 4 gets weighted more than all the other seasons combined because the denomonator is 8.

The first question: I do not account for strenght of opponents faced in the original DNRA, however, that is in the plans for the new version. Not that that does you much good right now. The old system used seasonal summaries of the PBP database and PCA was going to adjust the total expected run scoring against with the F/S Matrix run value for the opposing teams for each starting pitcher...I am probably going to shift to a system which totals up a pitcher's plate appearances against each team, figures out the ratios each team's offense produced (per PA) in the relevant statistics and essentially calculates a DNRA for each pitcher vs. team match-up. But that's a ways off still.

The second question: to get career DNRA+ is use the player's original defense neutral runs allowed and his original expected runs allowed...sum them all up and recalculate the + version on the career totals. That's, I believe the best way to do it.

[brett]

The second question: to get career DNRA+ is use the player's original defense neutral runs allowed and his original expected runs allowed...sum them all up and recalculate the + version on the career totals. That's, I believe the best way to do it.

My only point with this is that the higher run environment seasons will "accidentally" get weighted more.

Just compare 2 hypotheticals.

Hypothetical #1: 2 seasons.
Year 1:
Innings 200
Defense neutral runs allowed: 80
Expected: 100

Year 2:
Innings 200
Defense neutral runs allowed: 120
Expected: 100

This guy has given up 200 neutralized runs out of 200 expected for a DNRA+ of 100

Hypothetical #2
Year 1:
Innings 200
Defense neutral runs allowed: 80
Expected: 100

Year 2
IP: 200
Defense neutral runs allowed: 180
Expected: 150 (basically the run environment was 50% higher for whatever reason).

Now you get 260 defense neutral runs allowed and 250 expected. The DNRA+ is now 96 and the only reason is that the higher run environment of Year 2 makes it get weighted 3/2 as much as season 1.

The net practical result is that when a player has his best years in a higher run setting, they count more and he gets overrated. When he has his worst in a higher run setting, they count more and he gets underrated.

The other way to do it is take the relative ERA's of 0.8 and 1.2 for seasons 1 and 2 and weight them by innings (or equivalent innings) and then divide. This counts seasons of all run setting equally.

(1.2 x 200+0.8x 200)/200

[SABR Matt]

Huh.

That's actually a great point.

I hadn't thought of that...but you're correct. The way to account for that is to use the odds ratio method and set everyone's run environments to the same score. That's a relatively quick fix...so let me do that for my DIPItchers table and post new career figures.

[BigRon]

I'm no mathematician, but I think you guys are entering the world of harmonic means versus arithmetic means, and things like that. As Brett implied, you've got to be very careful when doing these things.

I think the old fuel economy example is illustrative. You own 2 cars. Car A gets 30 miles per gallon (MPG) and car B gets 20 MPG. You drive each 1000 miles per month. Your average fuel economy is 25 MPG, right? No, it's 24.0 MPG. Car A uses 33.3 gallons to travel 1000 miles, car B uses 50 gallons. Together, they consume 83.3 gallons to travel 2000 miles= 24.0 MPG.

To get an average of 25.0 MPG, Car A (which consumes less, or under-consumes) must travel 1500 miles while car B travels 1000 miles. Then, car A will use 50 gallons, car B will use 50 gallons, and the combined distance travelled with 100 gallons of fuel will be 2500 miles= 25.0 MPG.

There are lots of related things like looking at inverses incorectly, too. As I said, I'm not a math guy and don't even know many of the terms, but it is really easy to reach the wrong answer/conclusion if you don't follow the correct logic.

[SABR Matt]

The solution, as I said before, is to take every player's season and adjust it so that it appears they all played with the same run-scoring context. The way to do it is by assuming their relative production rates are constant in any environment (if you allow 10% more runs than you should in a 5 R/G league - resulting in a 5.50 DNRA...you're still going to allow 10% more runs than you should in a 4 R/G league, resulting in a 4.40 DNRA).

I reduced all pitching seasons to a 4.75 R/G context and recalculated my career figures for total neutral runs allowed and expected runs allowed. Here is the new DNRA+ leaderboard for pitchers with at least 3000 defense independent outs:

First Last Outs XRA NRA DNRA+ Marker
Pedro Martinez 6893 1212.66 781.23 155.2 197.8
Roger Clemens 13666 2404.20 1648.92 145.8 309.4
Randy Johnson 10075 1772.45 1272.40 139.3 225.7
Sandy Koufax 6684 1175.89 888.58 132.3 115.9
Greg Maddux 12525 2203.47 1666.32 132.2 247.6
Mike Mussina 8394 1476.72 1124.43 131.3 141.9
Kevin Brown 9036 1589.67 1231.38 129.1 157.5
Curt Schilling 8183 1439.60 1122.58 128.2 134.7
Bruce Sutter 3109 546.95 431.75 126.7 57.1
Kent Tekulve 4327 761.23 601.36 126.6 68.4
Hoyt Wilhelm 4853 853.77 676.96 126.1 73.0
Lee Smith 3892 684.70 544.04 125.9 61.1
Andy Pettitte 5737 1009.29 807.82 124.9 84.3
Tim Hudson 3846 676.61 542.24 124.8 56.4
Kerry Wood 3371 593.05 477.00 124.3 47.7
John Smoltz 8228 1447.52 1170.41 123.7 122.0
Stu Miller 3778 664.65 538.10 123.5 55.7
Rich Gossage 5390 948.24 771.59 122.9 94.5
Dan Quisenberry 3091 543.79 445.07 122.2 46.2
Bret Saberhagen 7314 1286.72 1054.26 122.0 100.4
Nolan Ryan 16256 2859.85 2350.79 121.7 217.2
J.R. Richard 4826 849.02 700.52 121.2 65.5
Tom Seaver 14283 2512.75 2082.87 120.6 187.1
Barry Zito 3577 629.29 523.74 120.2 49.1
Mark Buehrle 3611 635.27 528.87 120.1 45.2
Sam McDowell 7560 1330.00 1108.84 119.9 102.3
Tom Gordon 5914 1040.43 870.54 119.5 87.8
Bartolo Colon 4859 854.82 715.43 119.5 62.8
Dave Stieb 8650 1521.76 1276.23 119.2 109.5
Sid Fernandez 5613 987.47 832.36 118.6 68.6
Ron Guidry 7124 1253.30 1056.74 118.6 91.9
Juan Marichal 10576 1860.59 1569.13 118.6 134.1
Doug Jones 3059 538.16 454.02 118.5 47.6
Alej. Pena 3210 564.72 478.44 118.0 41.3
Jesse Orosco 3777 664.47 566.15 117.4 50.2
Andy Messersmith 6643 1168.68 995.78 117.4 76.3
John Franco 3602 633.69 539.93 117.4 46.1
Derek Lowe 3616 636.15 542.33 117.3 46.6
Dwight Gooden 8118 1428.17 1217.67 117.3 98.9
Dean Chance 6425 1130.32 964.77 117.2 76.3
Mike Jackson 3323 584.60 499.13 117.1 48.4
Bob Gibson 11601 2040.92 1744.82 117.0 138.5
John Hiller 3722 654.80 560.03 116.9 48.8
Kevin Appier 7122 1252.94 1072.15 116.9 91.8
Gary Lavelle 3249 571.58 489.45 116.8 40.5
Whitey Ford 6420 1129.44 967.29 116.8 71.9
Gaylord Perry 16021 2818.51 2415.48 116.7 186.1
Bob Stanley 5151 906.19 777.22 116.6 60.4
Dave Righetti 4205 739.77 634.68 116.6 55.5
Rollie Fingers 5029 884.73 759.23 116.5 63.8
Larry Andersen 3003 528.31 453.78 116.4 41.6
Dick Hall 3252 572.11 491.98 116.3 40.6
Greg Minton 3391 596.56 513.01 116.3 47.5
Dan Plesac 3083 542.38 466.49 116.3 38.8
Jimmy Key 7714 1357.09 1168.19 116.2 89.3
Camilo Pascual 7545 1327.36 1143.84 116.0 90.9
David Cone 8148 1433.44 1239.56 115.6 97.2
Don Drysdale 9994 1758.20 1525.62 115.2 107.5
Bert Blyleven 14916 2624.11 2281.60 115.0 166.6
Sparky Lyle 4136 727.63 633.65 114.8 47.2
Jose Rijo 5655 994.86 866.62 114.8 68.5
Tug McGraw 4486 789.20 688.75 114.6 57.8
Dennis Eckersley 9867 1735.86 1518.29 114.3 118.5
Teddy Higuera 4120 724.81 634.23 114.3 49.6
Juan Guzman 3892 684.70 599.27 114.3 46.4
Orel Hershiser 8888 1563.63 1368.54 114.3 96.6
Jim Maloney 5525 971.99 854.76 113.7 61.0
Don Wilson 5186 912.35 803.54 113.5 51.4
Tom Glavine 11132 1958.41 1726.68 113.4 119.2
Ben McDonald 3832 674.15 595.21 113.3 39.6
Fergie Jenkins 13506 2376.06 2099.60 113.2 135.9
Jon Matlack 7111 1251.01 1107.05 113.0 73.2
Larry Jackson 8983 1580.34 1399.16 112.9 88.4
Don McMahon 3883 683.12 605.68 112.8 47.8
Chuck Finley 8934 1571.72 1396.31 112.6 89.3
Sam Jones 3526 620.31 551.25 112.5 34.3
Willie Hernandez 3120 548.89 489.42 112.2 38.0
Luis Tiant 10478 1843.35 1646.81 111.9 105.1
David Wells 8934 1571.72 1404.35 111.9 88.7
Steve Rogers 8449 1486.40 1328.15 111.9 83.8
Mark Langston 8732 1536.19 1373.38 111.9 88.0
Clay Carroll 3927 690.86 619.77 111.5 41.0
Gene Garber 4516 794.48 712.83 111.5 51.4
Steve Carlton 15702 2762.39 2478.78 111.4 161.4
Roger McDowell 3155 555.05 498.11 111.4 31.4
Kenny Rogers 8093 1423.77 1279.11 111.3 80.7
Bill Hands 5888 1035.85 931.45 111.2 55.2
Rick Reuschel 10726 1886.98 1701.42 110.9 105.8
Hank Aguirre 3909 687.69 620.46 110.8 41.7
Jim Palmer 11620 2044.26 1845.97 110.7 108.2
Terry Forster 3308 581.96 526.22 110.6 34.0
Bill Swift 4803 844.97 764.64 110.5 50.0
Gary Nolan 4996 878.93 795.56 110.5 44.9
Don Mossi 3882 682.94 618.64 110.4 34.6
Jim Bunning 10911 1919.53 1739.60 110.3 103.7
Wilson Alvarez 4723 830.90 753.33 110.3 46.7
Al Leiter 6523 1147.56 1040.81 110.3 64.3
Jeff Fassero 5620 988.70 898.77 110.0 56.8
Ron Perranoski 3502 616.09 560.15 110.0 37.1
Don Aase 3345 588.47 535.21 110.0 32.2
Greg Harris 4433 779.88 710.12 109.8 46.7
Mark Mulder 3561 626.47 570.67 109.8 33.4
John Tudor 5342 939.80 857.95 109.5 50.3
Jim O'Toole 4854 853.94 779.86 109.5 45.1
Matt Clement 3523 619.79 566.89 109.3 33.8
Javier Vazquez 4485 789.03 721.95 109.3 42.2
Warren Spahn 6587 1158.82 1060.91 109.2 57.8
Billy O'Dell 5316 935.22 856.73 109.2 47.0
Mel Stottlemyre 7928 1394.74 1279.45 109.0 66.7
Alex Fernandez 4797 843.92 774.27 109.0 42.9
Sonny Siebert 6534 1149.50 1054.96 109.0 58.2
Tommy John 14093 2479.32 2275.88 108.9 126.0
Chan Ho Park 4264 750.15 688.66 108.9 38.5
Chris Bosio 5140 904.26 830.98 108.8 44.6
Vida Blue 9963 1752.75 1611.28 108.8 96.1
Frank Viola 8535 1501.53 1382.26 108.6 77.9
Jason Schmidt 4573 804.51 741.26 108.5 44.8
Jeff Reardon 3368 592.52 546.00 108.5 30.7
Mario Soto 5184 912.00 841.33 108.4 48.3
Don Sutton 15676 2757.81 2544.26 108.4 131.5
Wilbur Wood 8088 1422.89 1313.87 108.3 71.5
Don Gullett 4081 717.95 663.74 108.2 33.3
Len Barker 4024 707.93 655.68 108.0 36.3
Larry Dierker 6945 1221.81 1131.91 107.9 60.7
Rudy May 7795 1371.34 1271.19 107.9 69.1
Jim Brewer 3067 539.56 500.49 107.8 33.0
John Denny 6488 1141.41 1058.79 107.8 59.2
Jon Lieber 5130 902.50 838.32 107.7 41.3
Jeff Russell 3327 585.31 543.81 107.6 33.3
Zane Smith 5744 1010.52 939.87 107.5 47.7
Ken Forsch 6285 1105.69 1028.41 107.5 51.8
Ron Reed 7345 1292.18 1202.66 107.4 60.9
Rick Aguilera 3651 642.31 597.84 107.4 33.4
Dave Rozema 3287 578.27 539.15 107.3 26.0
John Candelaria 7528 1324.37 1234.93 107.2 62.5
Brad Radke 6158 1083.35 1010.31 107.2 48.0
Atlee Hammaker 3246 571.06 533.43 107.1 27.9
Mike Witt 6278 1104.46 1032.12 107.0 52.2
Eric Plunk 3244 570.70 533.38 107.0 31.5
Pascual Perez 3718 654.09 611.46 107.0 35.8
Steve Bedrosian 3588 631.22 590.42 106.9 30.4
Mike Boddicker 6379 1122.23 1049.70 106.9 52.1
Frank Lary 4906 863.09 807.49 106.9 40.2
Bob Veale 5336 938.74 880.01 106.7 45.0
Bob Ojeda 5710 1004.54 941.98 106.6 48.8
Jerry Koosman 11437 2012.06 1887.16 106.6 88.6
Randy Jones 5786 1017.91 955.27 106.6 49.5
Burt Hooton 7936 1396.15 1310.64 106.5 64.1
Phil Niekro 16148 2840.85 2668.06 106.5 126.2
Bill Campbell 3659 643.71 604.65 106.5 30.1
Britt Burns 3255 572.64 538.30 106.4 25.2
Bob Miller 4650 818.06 769.10 106.4 42.2
Matt Morris 4060 714.26 671.53 106.4 31.7
Tim Belcher 6947 1222.16 1150.11 106.3 52.8
Mickey Lolich 10986 1932.72 1819.30 106.2 82.1
Ramon Martinez 5637 991.69 933.84 106.2 44.4
Jose DeLeon 5671 997.68 939.61 106.2 45.5
Tom Candiotti 8026 1411.98 1330.22 106.1 60.1
Mike Cuellar 8214 1445.06 1362.49 106.1 71.0
Billy Pierce 4196 738.19 696.13 106.0 30.5
Freddy Garcia 3606 634.39 598.34 106.0 27.3
Tom Burgmeier 3714 653.39 616.66 106.0 36.1
Mike Marshall 4061 714.44 674.60 105.9 35.0
Bill Singer 6519 1146.86 1083.27 105.9 50.4
Erik Hanson 4631 814.71 770.22 105.8 37.3
Lindy McDaniel 5970 1050.28 994.36 105.6 52.3
Frank Tanana 12610 2218.43 2100.93 105.6 96.3
Chris Carpenter 3403 598.68 567.25 105.5 26.8
Dave LaRoche 3111 547.31 518.67 105.5 27.3
Bill Monbouquette 5947 1046.23 993.34 105.3 42.6
Doug Drabek 7584 1334.22 1266.78 105.3 56.6
Ralph Terry 5386 947.54 900.83 105.2 40.5
Pat Hentgen 5626 989.76 941.17 105.2 45.6
Joe Horlen 5968 1049.93 1000.56 104.9 44.8
Mark Gubicza 6646 1169.20 1114.49 104.9 48.7
Geoff Zahn 5520 971.11 926.56 104.8 41.8
Rick Honeycutt 6510 1145.28 1093.03 104.8 52.8
John Smiley 5672 997.85 953.00 104.7 41.1
Jim Perry 9870 1736.39 1658.57 104.7 71.3
Jack Morris 11408 2006.96 1917.67 104.7 79.7
Cory Lidle 3436 604.48 578.03 104.6 25.2
Ed Halicki 3233 568.77 544.35 104.5 23.9
Jack Kralick 3652 642.48 615.03 104.5 27.1
Turk Farrell 5085 894.58 856.76 104.4 42.9
Pat Jarvis 3596 632.63 606.39 104.3 28.8
Bob Friend 6734 1184.69 1136.43 104.2 46.5
Jack Sanford 6154 1082.65 1038.69 104.2 42.3
Milt Pappas 9422 1657.57 1590.52 104.2 65.7
Greg Swindell 6517 1146.51 1100.64 104.2 48.2
Jose Guzman 3692 649.52 623.97 104.1 25.3
Charles Nagy 5278 928.54 892.33 104.1 44.0
Stan Williams 5306 933.46 897.12 104.1 36.5
Jack McDowell 5534 973.57 937.52 103.8 38.5
Ken Johnson 4966 873.65 841.42 103.8 35.3
Bryn Smith 5352 941.56 906.86 103.8 37.4
Bob Welch 9245 1626.44 1567.04 103.8 64.5
Jim Gott 3371 593.05 571.47 103.8 23.7
Jeff Weaver 3688 648.81 625.25 103.8 24.8
Ray Herbert 4691 825.27 796.00 103.7 31.9
Pete Richert 3480 612.22 590.62 103.7 25.4
Bruce Hurst 7302 1284.61 1239.66 103.6 51.3
Kevin Tapani 6376 1121.70 1082.58 103.6 42.0
Steve Barber 5968 1049.93 1013.39 103.6 44.1
Jim Abbott 4723 830.90 802.84 103.5 32.7
Tim Wakefield 6520 1147.04 1108.45 103.5 46.7
Shane Reynolds 4693 825.62 798.13 103.4 36.3
Jim Beattie 3459 608.53 589.04 103.3 27.6
Kevin Millwood 3951 695.08 672.91 103.3 25.0
Al Downing 6750 1187.50 1150.14 103.2 47.3
John Thomson 3362 591.46 573.22 103.2 22.4
Jarrod Washburn 3236 569.30 551.97 103.1 20.1
Harvey Haddix 3608 634.74 616.13 103.0 25.1
Danny Darwin 9019 1586.68 1540.88 103.0 65.2
Jim Barr 6217 1093.73 1062.25 103.0 39.7
Dick Donovan 4541 798.88 776.10 102.9 30.8
Steve Busby 3176 558.74 543.21 102.9 22.4
Rick Reed 4154 730.80 712.41 102.6 29.3
Earl Wilson 6254 1100.24 1073.46 102.5 38.5
Dave Dravecky 3116 548.19 535.05 102.5 18.5
Dennis Martinez 11939 2100.38 2050.28 102.4 85.4
Dave Goltz 6123 1077.19 1051.64 102.4 40.8
Dennis Lamp 5489 965.66 942.93 102.4 38.9
La Marr Hoyt 3909 687.69 671.64 102.4 26.7
Fritz Peterson 6594 1160.06 1133.29 102.4 43.4
Andy Benes 6904 1214.59 1186.73 102.3 45.3
Ray Culp 5786 1017.91 995.35 102.3 39.4
Jamie Moyer 8699 1530.38 1496.48 102.3 61.2
Chris Short 6958 1224.09 1197.16 102.2 50.1
Ismael Valdez 4868 856.41 838.30 102.2 36.2
Fernan. Valenzuela 8763 1541.64 1509.26 102.1 66.7
Jerry Reuss 10994 1934.13 1894.21 102.1 72.2
David Palmer 3275 576.16 564.27 102.1 20.2
Early Wynn 3967 697.90 683.73 102.1 23.6
Steve Avery 4384 771.26 756.27 102.0 27.6
Kirk McCaskill 5174 910.24 892.55 102.0 33.6
Dave Boswell 3210 564.72 553.89 102.0 23.6
Scott Erickson 6360 1118.89 1097.45 102.0 43.5
Bill Bonham 4532 797.30 782.81 101.9 28.5
Storm Davis 5328 937.33 920.75 101.8 35.3
Rick Waits 4314 758.94 746.61 101.7 29.1
Jim Bibby 5183 911.82 897.22 101.6 35.4
Johnny Podres 5520 971.11 955.61 101.6 34.7
Marty Pattin 6111 1075.08 1058.04 101.6 36.4
Mark Portugal 5009 881.21 867.39 101.6 31.1
Joey Hamilton 3729 656.03 645.78 101.6 26.6
Gary Bell 6041 1062.77 1046.35 101.6 37.0
Jim Kaat 13638 2399.28 2362.22 101.6 85.9
Catfish Hunter 10176 1790.22 1763.00 101.5 64.0
Rheal Cormier 3292 579.15 570.51 101.5 24.2
Mike Scott 6182 1087.57 1071.79 101.5 46.7
Ken Brett 4613 811.55 800.11 101.4 30.1
Craig Lefferts 3396 597.44 589.91 101.3 21.0
Mike Norris 3394 597.09 589.74 101.2 24.3
Dave Giusti 4944 869.78 859.19 101.2 29.0
Mike Hampton 5555 977.27 965.48 101.2 37.6
Dock Ellis 6303 1108.86 1095.52 101.2 40.7
Matt Young 3610 635.09 627.59 101.2 22.1
John Burkett 7474 1314.87 1299.52 101.2 43.6
Steve Stone 5393 948.77 938.45 101.1 30.8
Lee Stange 3654 642.83 635.85 101.1 21.0
Oil Can Boyd 4204 739.59 731.67 101.1 25.2
Skip Lockwood 3693 649.69 643.28 101.0 24.9
Hideo Nomo 5349 941.03 932.56 100.9 35.9
Mike Moore 8541 1502.58 1489.11 100.9 56.9
Jim Rooker 5424 954.22 945.87 100.9 32.3
Bruce Kison 5424 954.22 946.33 100.8 30.5
Moose Haas 4937 868.55 861.95 100.8 29.2
Woody Williams 5004 880.33 873.84 100.7 31.5
Rick Wise 9412 1655.81 1643.89 100.7 53.5
Denny McLain 5645 993.10 985.97 100.7 39.0
Charlie Leibrandt 6913 1216.18 1208.18 100.7 42.0
Moe Drabowsky 4655 818.94 813.68 100.6 35.1
Dennis Leonard 6505 1144.40 1137.35 100.6 38.1
Mike Caldwell 7217 1269.66 1262.41 100.6 45.5
Paul Foytack 3575 628.94 626.42 100.4 20.0
Chris Hammond 3279 576.86 574.69 100.4 22.3
Joe Nuxhall 4352 765.63 762.75 100.4 24.9
Joey Jay 4444 781.81 779.56 100.3 26.9
Claude Osteen 10267 1806.23 1801.20 100.3 56.2
Darold Knowles 3243 570.53 569.09 100.3 21.0
Dick Ellsworth 6418 1129.09 1126.62 100.2 38.9
Jose Mesa 4047 711.97 710.91 100.1 32.5
Shawn Estes 4282 753.31 752.47 100.1 26.1
Denny Neagle 5324 936.63 935.76 100.1 33.3
Bud Black 6113 1075.44 1074.83 100.1 34.5
Julian Tavarez 3082 542.20 542.21 100.0 20.9
Gene Conley 3190 561.20 561.99 99.9 18.7
Pedro Ramos 6269 1102.88 1105.06 99.8 38.1
Dave Roberts 6270 1103.06 1105.36 99.8 39.2
Larry Christenson 4194 737.83 739.54 99.8 23.7
Mike Flanagan 8177 1438.55 1442.48 99.7 45.8
Ed Figueroa 3886 683.65 685.66 99.7 21.8
Miguel Batista 3122 549.24 551.04 99.7 17.9
Bob Purkey 5646 993.28 996.78 99.6 30.7
Robin Roberts 6207 1091.97 1096.40 99.6 35.9
Melido Perez 4034 709.69 712.62 99.6 24.9
Bob Tewksbury 5361 943.14 947.13 99.6 28.2
Rick Helling 3849 677.14 680.21 99.5 25.4
Reggie Cleveland 5472 962.67 967.08 99.5 29.6
Tom Bradley 3093 544.14 546.90 99.5 19.0
Bobby Bolin 4753 836.18 840.45 99.5 29.3
Lew Burdette 5874 1033.39 1039.25 99.4 32.2
Gary Peters 6185 1088.10 1094.57 99.4 37.3
Bob Moose 3792 667.11 671.13 99.4 22.5
Dave McNally 8056 1417.26 1427.97 99.3 42.6
Mike Morgan 7877 1385.77 1396.77 99.2 47.6
Curt Simmons 5399 949.82 957.57 99.2 32.3
Dan Petry 6188 1088.63 1097.67 99.2 35.4
Pat Dobson 6317 1111.32 1120.88 99.1 32.5
Darryl Kile 5890 1036.20 1045.61 99.1 31.2
Vern Law 5203 915.34 923.72 99.1 29.5
Steve Trout 4541 798.88 806.51 99.1 29.4
Rick Rhoden 7758 1364.83 1378.77 99.0 40.3
Tom Underwood 4731 832.31 840.96 99.0 25.4
Grant Jackson 4032 709.33 717.20 98.9 25.8
Dick Tidrow 5213 917.10 927.44 98.9 29.3
Juan Berenguer 3593 632.10 640.30 98.7 22.0
Dick Drago 5607 986.42 999.36 98.7 27.9
Andy Ashby 4805 845.32 857.21 98.6 29.3
Scott Sanderson 7687 1352.34 1372.16 98.6 41.2
Scott McGregor 6344 1116.07 1132.73 98.5 31.1
Dave Burba 4686 824.39 837.43 98.4 23.8
Pedro Astacio 5644 992.93 1009.39 98.4 29.3
Charlie Hough 11411 2007.49 2043.54 98.2 59.2
Esteban Loaiza 5278 928.54 946.32 98.1 29.2
Jim Merritt 4456 783.93 799.16 98.1 22.4
Doc Medich 5957 1047.99 1068.37 98.1 26.8
Steve Hargan 4927 866.79 883.73 98.1 34.2
Danny Jackson 6220 1094.26 1115.75 98.1 35.6
Sidney Ponson 3695 650.05 662.88 98.1 19.1
Dave Wickersham 3384 595.33 607.30 98.0 16.6
Russ Ortiz 3716 653.74 666.90 98.0 21.1
Cal McLish 3843 676.08 689.72 98.0 20.2
John Montefusco 4958 872.24 891.19 97.9 23.4
Vern Ruhle 4231 744.34 760.56 97.9 21.1
Ken Hill 5493 966.36 987.43 97.9 30.7
Dustin Hermanson 3167 557.16 569.47 97.8 18.8
Todd Stottlemyre 6278 1104.46 1129.02 97.8 29.0
Joe Magrane 3289 578.62 591.55 97.8 16.6
Orlando Pena 3624 637.56 652.11 97.8 19.3
John Buzhardt 4382 770.91 788.66 97.7 23.4
Pete Vuckovich 4318 759.65 777.15 97.7 20.5
Ron Darling 7097 1248.55 1277.47 97.7 32.6
Dave Stewart 7840 1379.26 1411.37 97.7 37.7
Shane Rawley 5631 990.64 1014.26 97.7 28.5
Al Fitzmorris 3797 667.99 684.23 97.6 18.3
Rick Sutcliffe 8136 1431.33 1466.22 97.6 44.7
Woodie Fryman 7175 1262.27 1293.58 97.6 38.7
Livan Hernandez 5203 915.34 938.81 97.5 25.5
Lary Sorensen 5203 915.34 939.30 97.4 24.0
Bill Lee 5876 1033.74 1060.99 97.4 27.7
Pete Harnisch 5267 926.60 951.46 97.4 26.8
Paul Quantrill 3565 627.18 644.04 97.4 22.2
Joe Coleman 7723 1358.68 1396.30 97.3 40.0
Dick Ruthven 6386 1123.46 1154.69 97.3 28.1
Denny Lemaster 5048 888.07 912.99 97.3 21.3
Wally Bunker 3227 567.71 584.18 97.2 16.4
Charles Hudson 3054 537.28 552.98 97.2 14.7
Ken Holtzman 8560 1505.93 1549.97 97.2 40.8
Steve Trachsel 5822 1024.24 1054.41 97.1 24.1
Bob Turley 3011 529.71 545.38 97.1 14.3
Jim Bouton 3697 650.40 669.88 97.1 18.1
Steve Renko 7484 1316.63 1356.77 97.0 32.3
Roger Craig 3772 663.59 683.88 97.0 21.1
Joaquin Andujar 6381 1122.58 1157.25 97.0 28.6
Pedro Borbon 3027 532.53 549.54 96.9 14.5
Cal Eldred 3779 664.82 686.13 96.9 17.7
Rick Langford 4489 789.73 815.61 96.8 22.2
Hal Brown 3201 563.14 581.63 96.8 14.3
Jim Clancy 7532 1325.07 1371.52 96.6 34.8
Ernie Broglio 3959 696.49 721.24 96.6 21.7
Jack Billingham 6605 1161.99 1204.17 96.5 30.4
Ted Power 3472 610.81 633.03 96.5 17.7
Terry Mulholland 7225 1271.06 1318.44 96.4 35.6
Mudcat Grant 7297 1283.73 1331.75 96.4 32.2
Jim Colborn 4791 842.86 874.90 96.3 20.2
Mark Clark 3514 618.20 641.78 96.3 16.3
Floyd Bannister 7154 1258.57 1308.15 96.2 30.4
Diego Segui 5391 948.42 985.96 96.2 22.5
Kent Mercker 3466 609.76 633.90 96.2 18.2
Mike McCormick 7060 1242.04 1291.29 96.2 29.5
Fred Norman 5766 1014.39 1054.80 96.2 25.6
Bob McClure 3455 607.82 632.09 96.2 18.4
Don Cardwell 6065 1066.99 1109.67 96.2 25.1
Juan Pizarro 5960 1048.52 1090.75 96.1 29.9
Larry Gura 6051 1064.53 1109.20 96.0 27.0
Frank Castillo 4766 838.46 873.66 96.0 21.9
Kirk Rueter 5146 905.31 944.36 95.9 23.5
Wayne Garland 3112 547.48 571.14 95.9 15.9
Doyle Alexander 9989 1757.32 1835.31 95.8 45.6
Carl Morton 4862 855.35 893.34 95.7 19.9
Paul Lindblad 3602 633.69 661.91 95.7 18.2
Walt Terrell 5927 1042.71 1089.64 95.7 23.0
Bob Shaw 5230 920.09 961.72 95.7 24.0
Jim Lonborg 7448 1310.30 1369.96 95.6 30.2
Bob Bruce 3384 595.33 622.53 95.6 14.7
Bill Gullickson 7644 1344.78 1406.74 95.6 30.1
Clyde Wright 5142 904.61 946.77 95.5 20.5
Bob Shirley 4267 750.68 785.68 95.5 17.2
Pat Zachry 3535 621.90 651.45 95.5 14.9
Omar Olivares 4158 731.50 767.10 95.4 17.5
Aaron Sele 5354 941.91 987.90 95.3 23.0
Stan Bahnsen 7606 1338.09 1403.94 95.3 29.9
Bruce Ruffin 3844 676.26 709.88 95.3 25.0
Joe Gibbon 3224 567.19 595.77 95.2 14.9
Bill Wegman 4424 778.30 817.60 95.2 18.0
Mark Leiter 3558 625.94 658.21 95.1 14.1
Mike Krukow 6620 1164.63 1225.09 95.1 23.3
Rick Mahler 5859 1030.75 1085.46 95.0 20.4
Mark Davis 3460 608.70 641.31 94.9 23.4
Don Robinson 5852 1029.52 1085.39 94.9 23.0
Dick Bosman 4759 837.23 882.83 94.8 22.7
Craig Swan 3692 649.52 685.38 94.8 19.1
Bob Buhl 5511 969.53 1023.06 94.8 21.4
Tom Browning 5770 1015.09 1071.63 94.7 20.1
Paul Splittorff 7589 1335.10 1410.36 94.7 27.9
Tom Hume 3258 573.17 605.71 94.6 16.0
Eddie Fisher 4546 799.76 846.26 94.5 19.8
Bobby Witt 6901 1214.06 1284.76 94.5 24.5
Phil Regan 4128 726.22 768.72 94.5 19.7
Ed Whitson 6693 1177.47 1247.11 94.4 26.2
Dan Schatzeder 3924 690.33 731.47 94.4 17.5
Mike Torrez 9075 1596.53 1691.75 94.4 34.5
Nelson Briles 6311 1110.27 1176.97 94.3 22.6
Ray Washburn 3592 631.93 670.82 94.2 14.4
Andy Hassler 3361 591.29 627.73 94.2 15.6
Kevin Gross 7516 1322.26 1404.61 94.1 24.0
Milt Wilcox 5964 1049.22 1114.69 94.1 20.8
Ron Kline 4557 801.69 851.95 94.1 17.5
Bobby Jones 4354 765.98 814.30 94.1 14.5
Larry McWilliams 4681 823.51 876.92 93.9 20.0
Jim Deshaies 4569 803.81 856.76 93.8 17.7
Glen Hobbie 3752 660.07 704.45 93.7 12.7
Danny Cox 3829 673.62 719.26 93.7 17.2
Pete Schourek 3113 547.66 585.06 93.6 13.3
Doug Rau 3734 656.91 702.02 93.6 12.2
Mark Gardner 4856 854.30 913.16 93.6 14.2
Dan Spillner 4514 794.13 849.44 93.5 17.5
Al Jackson 4169 733.44 785.27 93.4 13.9
Eric Show 4915 864.68 926.76 93.3 16.4
Barry Latman 3631 638.79 684.73 93.3 16.1
Eric Milton 3520 619.26 664.58 93.2 11.2
Bob Knepper 8115 1427.64 1534.86 93.0 29.4
Doug Bird 3615 635.97 683.86 93.0 13.1
Steve Sparks 3505 616.62 663.50 92.9 10.9
Brett Tomko 3753 660.25 711.76 92.8 11.6
Steve McCatty 3581 629.99 679.37 92.7 12.2
Jim McGlothlin 3849 677.14 730.70 92.7 14.5
Steve Blass 4587 806.97 871.21 92.6 16.5
Mike Bielecki 3715 653.56 705.89 92.6 11.1
Ray Sadecki 7503 1319.97 1426.63 92.5 24.3
Gene Nelson 3208 564.37 610.03 92.5 13.0
Clay Kirby 4566 803.28 868.59 92.5 14.7
George Brunet 4220 742.41 803.60 92.4 16.7
Ster. Hitchcock 3252 572.11 619.58 92.3 8.4
Bob Forsch 8268 1454.56 1577.62 92.2 24.1
Jason Bere 3138 552.06 599.78 92.0 11.2
Brian Anderson 4233 744.69 810.06 91.9 10.4
Joe Niekro 10695 1881.53 2048.44 91.9 29.6
Glendon Rusch 3894 685.06 746.79 91.7 10.8
Bill Stoneman 3673 646.18 704.45 91.7 13.1
Sammy Ellis 3037 534.29 582.90 91.7 11.6
Dennis Rasmussen 4358 766.69 837.37 91.6 18.2
Richard Dotson 5529 972.69 1062.65 91.5 15.3
Tom Murphy 4264 750.15 819.83 91.5 14.1
Jason Johnson 3288 578.44 632.97 91.4 6.9
Tom Phoebus 3040 534.81 585.24 91.4 7.8
Tim Leary 4480 788.15 862.51 91.4 11.5
Mike LaCoss 5156 907.07 994.25 91.2 13.5
Lynn McGlothen 4499 791.49 867.98 91.2 13.3
Tony Cloninger 5005 880.51 966.02 91.1 12.6
Ramon Ortiz 3040 534.81 589.43 90.7 5.7
Darren Oliver 3677 646.88 713.27 90.7 9.5
Pat Rapp 3697 650.40 717.18 90.7 7.9
Willie Blair 3411 600.08 662.53 90.6 8.2
Bob Walk 4996 878.93 970.50 90.6 10.1
Curt Young 3304 581.26 642.95 90.4 9.0
Neal Heaton 4509 793.25 878.34 90.3 9.7
John Curtis 4929 867.14 960.59 90.3 10.0
Chuck Dobson 3763 662.01 733.41 90.3 7.3
Matt Keough 3589 631.40 700.01 90.2 8.8
Jaime Navarro 5664 996.44 1105.40 90.1 14.2
Jim Nash 3310 582.31 646.25 90.1 8.5
James Baldwin 3354 590.06 655.00 90.1 8.4
Bl. Mo. Odom 4504 792.37 882.02 89.8 13.3
Pete Smith 3091 543.79 605.46 89.8 6.2
Alan Foster 3054 537.28 598.79 89.7 8.8
Jim Slaton 7998 1407.06 1571.00 89.6 13.1
Jack Fisher 5892 1036.56 1158.35 89.5 10.6
Bill Krueger 3608 634.74 709.92 89.4 7.1
Eric Rasmussen 3009 529.36 592.52 89.3 6.7
Mike Smithson 4070 716.02 802.27 89.2 5.3
Glenn Abbott 3841 675.73 757.22 89.2 7.6
Bill Travers 3368 592.52 664.16 89.2 5.8
Jeff Suppan 4336 762.81 855.31 89.2 6.4
Wayne Twitchell 3153 554.69 623.47 89.0 9.9
Ray Burris 6579 1157.42 1305.81 88.6 11.7
Tom Griffin 4517 794.66 899.22 88.4 9.6
Jamey Wright 3674 646.35 733.92 88.1 4.2
George Stone 3018 530.94 606.17 87.6 4.6
Ross Grimsley 5928 1042.89 1194.10 87.3 6.9
Lew Krausse 3835 674.68 776.95 86.8 3.8
Jimmy Haynes 3151 554.34 639.28 86.7 5.5
Andy Hawkins 4625 813.66 938.39 86.7 3.2
Dave LaPoint 4419 777.42 896.82 86.7 5.6
Ricky Bones 3659 643.71 743.54 86.6 4.0
Jose Lima 3900 686.11 797.42 86.0 3.5
Jesse Jefferson 3287 578.27 672.13 86.0 3.5
Dave Mlicki 3079 541.68 646.35 83.8 -0.4
Pete Falcone 4256 748.74 899.40 83.2 0.0
Randy Lerch 3248 571.41 736.10 77.6 -8.1

[brett]

I'm no mathematician, but I think you guys are entering the world of harmonic means versus arithmetic means, and things like that. As Brett implied, you've got to be very careful when doing these things.

I think the old fuel economy example is illustrative. You own 2 cars. Car A gets 30 miles per gallon (MPG) and car B gets 20 MPG. You drive each 1000 miles per month. Your average fuel economy is 25 MPG, right? No, it's 24.0 MPG. Car A uses 33.3 gallons to travel 1000 miles, car B uses 50 gallons. Together, they consume 83.3 gallons to travel 2000 miles= 24.0 MPG.

To get an average of 25.0 MPG, Car A (which consumes less, or under-consumes) must travel 1500 miles while car B travels 1000 miles. Then, car A will use 50 gallons, car B will use 50 gallons, and the combined distance travelled with 100 gallons of fuel will be 2500 miles= 25.0 MPG.

There are lots of related things like looking at inverses incorectly, too. As I said, I'm not a math guy and don't even know many of the terms, but it is really easy to reach the wrong answer/conclusion if you don't follow the correct logic.

What you have described is the problem with the current method of taking those averages.

[brett]

The solution, as I said before, is to take every player's season and adjust it so that it appears they all played with the same run-scoring context. The way to do it is by assuming their relative production rates are constant in any environment (if you allow 10% more runs than you should in a 5 R/G league - resulting in a 5.50 DNRA...you're still going to allow 10% more runs than you should in a 4 R/G league, resulting in a 4.40 DNRA).

I reduced all pitching seasons to a 4.75 R/G context and recalculated my career figures for total neutral runs allowed and expected runs allowed. Here is the new DNRA+ leaderboard for pitchers with at least 3000 defense independent outs:

First Last Outs XRA NRA DNRA+ Marker
Pedro Martinez 6893 1212.66 781.23 155.2 197.8
etc.

What year is that data through?

Any major movers?

I notice that relievers tend to drop a lot, like 20 points or so versus traditional ERA+. Is that due to their situational advantage being negated?

Also with that in mind, I'm wondering about where Rivera would fit right now. What's your estimate?

Wilhelm and Gossage look good, with Tekulve next among relievers.

[SABR Matt]

It's through 2005 still...and 1999 is still missing...I just applied the corrected method to my old excel table.

Rivera has a corrected DNRA+ of 192 excluding 1999 and 2006-2009...so...I'm thinking he fits in as the greatest reliever of all time. :) He just didn't have 3000 outs yet without 99 and 2006 onward.

Actually...relievers dropped because I fixed a small bug in my table that I didn't notice before today (thank goodness I went back in there to make this other correction or I'd be using bad data...for which I apologize...feel rather dumb at the moment).

I was calculating my pythagorean pitchers' winning percentages with an incorrectly defined X value (remember...PythagenPat W% uses a variable exponent based on run scoring rate per game: X = ((RA + XRA)/(Outs/27))^0.285 for pitchers. FOr some reason I had it as X = ((XRA*54) / Outs)^0.285...in other words...I was doing pythags based on the average run scoring environment...not the pitcher's run scoring environment. You won't notice big changes for the starting pitchers (whose run environments tend not to differ all that much from average) but for relievers running 150 DNRA+ values in their good years, it does make a noticeable difference and has now been corrected both for seasonal DNRA+ and for the same-context DNRA+ used in career DNRA+ calculation.

[curveball]

The solution, as I said before, is to take every player's season and adjust it so that it appears they all played with the same run-scoring context. The way to do it is by assuming their relative production rates are constant in any environment (if you allow 10% more runs than you should in a 5 R/G league - resulting in a 5.50 DNRA...you're still going to allow 10% more runs than you should in a 4 R/G league, resulting in a 4.40 DNRA).

I reduced all pitching seasons to a 4.75 R/G context and recalculated my career figures for total neutral runs allowed and expected runs allowed. Here is the new DNRA+ leaderboard for pitchers with at least 3000 defense independent outs:

First Last Outs XRA NRA DNRA+ Marker
Pedro Martinez 6893 1212.66 781.23 155.2 197.8
Roger Clemens 13666 2404.20 1648.92 145.8 309.4
Randy Johnson 10075 1772.45 1272.40 139.3 225.7
Sandy Koufax 6684 1175.89 888.58 132.3 115.9
Greg Maddux 12525 2203.47 1666.32 132.2 247.6
Mike Mussina 8394 1476.72 1124.43 131.3 141.9
Kevin Brown 9036 1589.67 1231.38 129.1 157.5
Curt Schilling 8183 1439.60 1122.58 128.2 134.7
Bruce Sutter 3109 546.95 431.75 126.7 57.1
Kent Tekulve 4327 761.23 601.36 126.6 68.4
Hoyt Wilhelm 4853 853.77 676.96 126.1 73.0
Lee Smith 3892 684.70 544.04 125.9 61.1
Andy Pettitte 5737 1009.29 807.82 124.9 84.3
Tim Hudson 3846 676.61 542.24 124.8 56.4
Kerry Wood 3371 593.05 477.00 124.3 47.7
John Smoltz 8228 1447.52 1170.41 123.7 122.0
Stu Miller 3778 664.65 538.10 123.5 55.7
Rich Gossage 5390 948.24 771.59 122.9 94.5
Dan Quisenberry 3091 543.79 445.07 122.2 46.2
Bret Saberhagen 7314 1286.72 1054.26 122.0 100.4
Nolan Ryan 16256 2859.85 2350.79 121.7 217.2
J.R. Richard 4826 849.02 700.52 121.2 65.5
Tom Seaver 14283 2512.75 2082.87 120.6 187.1
Barry Zito 3577 629.29 523.74 120.2 49.1
Mark Buehrle 3611 635.27 528.87 120.1 45.2
Sam McDowell 7560 1330.00 1108.84 119.9 102.3
Tom Gordon 5914 1040.43 870.54 119.5 87.8
Bartolo Colon 4859 854.82 715.43 119.5 62.8
Dave Stieb 8650 1521.76 1276.23 119.2 109.5
Sid Fernandez 5613 987.47 832.36 118.6 68.6
Ron Guidry 7124 1253.30 1056.74 118.6 91.9
Juan Marichal 10576 1860.59 1569.13 118.6 134.1
Doug Jones 3059 538.16 454.02 118.5 47.6
Alej. Pena 3210 564.72 478.44 118.0 41.3
Jesse Orosco 3777 664.47 566.15 117.4 50.2
Andy Messersmith 6643 1168.68 995.78 117.4 76.3
John Franco 3602 633.69 539.93 117.4 46.1
Derek Lowe 3616 636.15 542.33 117.3 46.6
Dwight Gooden 8118 1428.17 1217.67 117.3 98.9
Dean Chance 6425 1130.32 964.77 117.2 76.3
Mike Jackson 3323 584.60 499.13 117.1 48.4
Bob Gibson 11601 2040.92 1744.82 117.0 138.5
John Hiller 3722 654.80 560.03 116.9 48.8
Kevin Appier 7122 1252.94 1072.15 116.9 91.8
Gary Lavelle 3249 571.58 489.45 116.8 40.5
Whitey Ford 6420 1129.44 967.29 116.8 71.9
Gaylord Perry 16021 2818.51 2415.48 116.7 186.1
Bob Stanley 5151 906.19 777.22 116.6 60.4
Dave Righetti 4205 739.77 634.68 116.6 55.5
Rollie Fingers 5029 884.73 759.23 116.5 63.8
Larry Andersen 3003 528.31 453.78 116.4 41.6
Dick Hall 3252 572.11 491.98 116.3 40.6
Greg Minton 3391 596.56 513.01 116.3 47.5
Dan Plesac 3083 542.38 466.49 116.3 38.8
Jimmy Key 7714 1357.09 1168.19 116.2 89.3
Camilo Pascual 7545 1327.36 1143.84 116.0 90.9
David Cone 8148 1433.44 1239.56 115.6 97.2
Don Drysdale 9994 1758.20 1525.62 115.2 107.5
Bert Blyleven 14916 2624.11 2281.60 115.0 166.6
Sparky Lyle 4136 727.63 633.65 114.8 47.2
Jose Rijo 5655 994.86 866.62 114.8 68.5
Tug McGraw 4486 789.20 688.75 114.6 57.8
Dennis Eckersley 9867 1735.86 1518.29 114.3 118.5
Teddy Higuera 4120 724.81 634.23 114.3 49.6
Juan Guzman 3892 684.70 599.27 114.3 46.4
Orel Hershiser 8888 1563.63 1368.54 114.3 96.6
Jim Maloney 5525 971.99 854.76 113.7 61.0
Don Wilson 5186 912.35 803.54 113.5 51.4
Tom Glavine 11132 1958.41 1726.68 113.4 119.2
Ben McDonald 3832 674.15 595.21 113.3 39.6
Fergie Jenkins 13506 2376.06 2099.60 113.2 135.9
Jon Matlack 7111 1251.01 1107.05 113.0 73.2
Larry Jackson 8983 1580.34 1399.16 112.9 88.4
Don McMahon 3883 683.12 605.68 112.8 47.8
Chuck Finley 8934 1571.72 1396.31 112.6 89.3
Sam Jones 3526 620.31 551.25 112.5 34.3
Willie Hernandez 3120 548.89 489.42 112.2 38.0
Luis Tiant 10478 1843.35 1646.81 111.9 105.1
David Wells 8934 1571.72 1404.35 111.9 88.7
Steve Rogers 8449 1486.40 1328.15 111.9 83.8
Mark Langston 8732 1536.19 1373.38 111.9 88.0
Clay Carroll 3927 690.86 619.77 111.5 41.0
Gene Garber 4516 794.48 712.83 111.5 51.4
Steve Carlton 15702 2762.39 2478.78 111.4 161.4
Roger McDowell 3155 555.05 498.11 111.4 31.4
Kenny Rogers 8093 1423.77 1279.11 111.3 80.7
Bill Hands 5888 1035.85 931.45 111.2 55.2
Rick Reuschel 10726 1886.98 1701.42 110.9 105.8
Hank Aguirre 3909 687.69 620.46 110.8 41.7
Jim Palmer 11620 2044.26 1845.97 110.7 108.2
Terry Forster 3308 581.96 526.22 110.6 34.0
Bill Swift 4803 844.97 764.64 110.5 50.0
Gary Nolan 4996 878.93 795.56 110.5 44.9
Don Mossi 3882 682.94 618.64 110.4 34.6
Jim Bunning 10911 1919.53 1739.60 110.3 103.7
Wilson Alvarez 4723 830.90 753.33 110.3 46.7
Al Leiter 6523 1147.56 1040.81 110.3 64.3
Jeff Fassero 5620 988.70 898.77 110.0 56.8
Ron Perranoski 3502 616.09 560.15 110.0 37.1
Don Aase 3345 588.47 535.21 110.0 32.2
Greg Harris 4433 779.88 710.12 109.8 46.7
Mark Mulder 3561 626.47 570.67 109.8 33.4
John Tudor 5342 939.80 857.95 109.5 50.3
Jim O'Toole 4854 853.94 779.86 109.5 45.1
Matt Clement 3523 619.79 566.89 109.3 33.8
Javier Vazquez 4485 789.03 721.95 109.3 42.2
Warren Spahn 6587 1158.82 1060.91 109.2 57.8
Billy O'Dell 5316 935.22 856.73 109.2 47.0
Mel Stottlemyre 7928 1394.74 1279.45 109.0 66.7
Alex Fernandez 4797 843.92 774.27 109.0 42.9
Sonny Siebert 6534 1149.50 1054.96 109.0 58.2
Tommy John 14093 2479.32 2275.88 108.9 126.0
Chan Ho Park 4264 750.15 688.66 108.9 38.5
Chris Bosio 5140 904.26 830.98 108.8 44.6
Vida Blue 9963 1752.75 1611.28 108.8 96.1
Frank Viola 8535 1501.53 1382.26 108.6 77.9
Jason Schmidt 4573 804.51 741.26 108.5 44.8
Jeff Reardon 3368 592.52 546.00 108.5 30.7
Mario Soto 5184 912.00 841.33 108.4 48.3
Don Sutton 15676 2757.81 2544.26 108.4 131.5
Wilbur Wood 8088 1422.89 1313.87 108.3 71.5
Don Gullett 4081 717.95 663.74 108.2 33.3
Len Barker 4024 707.93 655.68 108.0 36.3
Larry Dierker 6945 1221.81 1131.91 107.9 60.7
Rudy May 7795 1371.34 1271.19 107.9 69.1
Jim Brewer 3067 539.56 500.49 107.8 33.0
John Denny 6488 1141.41 1058.79 107.8 59.2
Jon Lieber 5130 902.50 838.32 107.7 41.3
Jeff Russell 3327 585.31 543.81 107.6 33.3
Zane Smith 5744 1010.52 939.87 107.5 47.7
Ken Forsch 6285 1105.69 1028.41 107.5 51.8
Ron Reed 7345 1292.18 1202.66 107.4 60.9
Rick Aguilera 3651 642.31 597.84 107.4 33.4
Dave Rozema 3287 578.27 539.15 107.3 26.0
John Candelaria 7528 1324.37 1234.93 107.2 62.5
Brad Radke 6158 1083.35 1010.31 107.2 48.0
Atlee Hammaker 3246 571.06 533.43 107.1 27.9
Mike Witt 6278 1104.46 1032.12 107.0 52.2
Eric Plunk 3244 570.70 533.38 107.0 31.5
Pascual Perez 3718 654.09 611.46 107.0 35.8
Steve Bedrosian 3588 631.22 590.42 106.9 30.4
Mike Boddicker 6379 1122.23 1049.70 106.9 52.1
Frank Lary 4906 863.09 807.49 106.9 40.2
Bob Veale 5336 938.74 880.01 106.7 45.0
Bob Ojeda 5710 1004.54 941.98 106.6 48.8
Jerry Koosman 11437 2012.06 1887.16 106.6 88.6
Randy Jones 5786 1017.91 955.27 106.6 49.5
Burt Hooton 7936 1396.15 1310.64 106.5 64.1
Phil Niekro 16148 2840.85 2668.06 106.5 126.2
Bill Campbell 3659 643.71 604.65 106.5 30.1
Britt Burns 3255 572.64 538.30 106.4 25.2
Bob Miller 4650 818.06 769.10 106.4 42.2
Matt Morris 4060 714.26 671.53 106.4 31.7
Tim Belcher 6947 1222.16 1150.11 106.3 52.8
Mickey Lolich 10986 1932.72 1819.30 106.2 82.1
Ramon Martinez 5637 991.69 933.84 106.2 44.4
Jose DeLeon 5671 997.68 939.61 106.2 45.5
Tom Candiotti 8026 1411.98 1330.22 106.1 60.1
Mike Cuellar 8214 1445.06 1362.49 106.1 71.0
Billy Pierce 4196 738.19 696.13 106.0 30.5
Freddy Garcia 3606 634.39 598.34 106.0 27.3
Tom Burgmeier 3714 653.39 616.66 106.0 36.1
Mike Marshall 4061 714.44 674.60 105.9 35.0
Bill Singer 6519 1146.86 1083.27 105.9 50.4
Erik Hanson 4631 814.71 770.22 105.8 37.3
Lindy McDaniel 5970 1050.28 994.36 105.6 52.3
Frank Tanana 12610 2218.43 2100.93 105.6 96.3
Chris Carpenter 3403 598.68 567.25 105.5 26.8
Dave LaRoche 3111 547.31 518.67 105.5 27.3
Bill Monbouquette 5947 1046.23 993.34 105.3 42.6
Doug Drabek 7584 1334.22 1266.78 105.3 56.6
Ralph Terry 5386 947.54 900.83 105.2 40.5
Pat Hentgen 5626 989.76 941.17 105.2 45.6
Joe Horlen 5968 1049.93 1000.56 104.9 44.8
Mark Gubicza 6646 1169.20 1114.49 104.9 48.7
Geoff Zahn 5520 971.11 926.56 104.8 41.8
Rick Honeycutt 6510 1145.28 1093.03 104.8 52.8
John Smiley 5672 997.85 953.00 104.7 41.1
Jim Perry 9870 1736.39 1658.57 104.7 71.3
Jack Morris 11408 2006.96 1917.67 104.7 79.7
Cory Lidle 3436 604.48 578.03 104.6 25.2
Ed Halicki 3233 568.77 544.35 104.5 23.9
Jack Kralick 3652 642.48 615.03 104.5 27.1
Turk Farrell 5085 894.58 856.76 104.4 42.9
Pat Jarvis 3596 632.63 606.39 104.3 28.8
Bob Friend 6734 1184.69 1136.43 104.2 46.5
Jack Sanford 6154 1082.65 1038.69 104.2 42.3
Milt Pappas 9422 1657.57 1590.52 104.2 65.7
Greg Swindell 6517 1146.51 1100.64 104.2 48.2
Jose Guzman 3692 649.52 623.97 104.1 25.3
Charles Nagy 5278 928.54 892.33 104.1 44.0
Stan Williams 5306 933.46 897.12 104.1 36.5
Jack McDowell 5534 973.57 937.52 103.8 38.5
Ken Johnson 4966 873.65 841.42 103.8 35.3
Bryn Smith 5352 941.56 906.86 103.8 37.4
Bob Welch 9245 1626.44 1567.04 103.8 64.5
Jim Gott 3371 593.05 571.47 103.8 23.7
Jeff Weaver 3688 648.81 625.25 103.8 24.8
Ray Herbert 4691 825.27 796.00 103.7 31.9
Pete Richert 3480 612.22 590.62 103.7 25.4
Bruce Hurst 7302 1284.61 1239.66 103.6 51.3
Kevin Tapani 6376 1121.70 1082.58 103.6 42.0
Steve Barber 5968 1049.93 1013.39 103.6 44.1
Jim Abbott 4723 830.90 802.84 103.5 32.7
Tim Wakefield 6520 1147.04 1108.45 103.5 46.7
Shane Reynolds 4693 825.62 798.13 103.4 36.3
Jim Beattie 3459 608.53 589.04 103.3 27.6
Kevin Millwood 3951 695.08 672.91 103.3 25.0
Al Downing 6750 1187.50 1150.14 103.2 47.3
John Thomson 3362 591.46 573.22 103.2 22.4
Jarrod Washburn 3236 569.30 551.97 103.1 20.1
Harvey Haddix 3608 634.74 616.13 103.0 25.1
Danny Darwin 9019 1586.68 1540.88 103.0 65.2
Jim Barr 6217 1093.73 1062.25 103.0 39.7
Dick Donovan 4541 798.88 776.10 102.9 30.8
Steve Busby 3176 558.74 543.21 102.9 22.4
Rick Reed 4154 730.80 712.41 102.6 29.3
Earl Wilson 6254 1100.24 1073.46 102.5 38.5
Dave Dravecky 3116 548.19 535.05 102.5 18.5
Dennis Martinez 11939 2100.38 2050.28 102.4 85.4
Dave Goltz 6123 1077.19 1051.64 102.4 40.8
Dennis Lamp 5489 965.66 942.93 102.4 38.9
La Marr Hoyt 3909 687.69 671.64 102.4 26.7
Fritz Peterson 6594 1160.06 1133.29 102.4 43.4
Andy Benes 6904 1214.59 1186.73 102.3 45.3
Ray Culp 5786 1017.91 995.35 102.3 39.4
Jamie Moyer 8699 1530.38 1496.48 102.3 61.2
Chris Short 6958 1224.09 1197.16 102.2 50.1
Ismael Valdez 4868 856.41 838.30 102.2 36.2
Fernan. Valenzuela 8763 1541.64 1509.26 102.1 66.7
Jerry Reuss 10994 1934.13 1894.21 102.1 72.2
David Palmer 3275 576.16 564.27 102.1 20.2
Early Wynn 3967 697.90 683.73 102.1 23.6
Steve Avery 4384 771.26 756.27 102.0 27.6
Kirk McCaskill 5174 910.24 892.55 102.0 33.6
Dave Boswell 3210 564.72 553.89 102.0 23.6
Scott Erickson 6360 1118.89 1097.45 102.0 43.5
Bill Bonham 4532 797.30 782.81 101.9 28.5
Storm Davis 5328 937.33 920.75 101.8 35.3
Rick Waits 4314 758.94 746.61 101.7 29.1
Jim Bibby 5183 911.82 897.22 101.6 35.4
Johnny Podres 5520 971.11 955.61 101.6 34.7
Marty Pattin 6111 1075.08 1058.04 101.6 36.4
Mark Portugal 5009 881.21 867.39 101.6 31.1
Joey Hamilton 3729 656.03 645.78 101.6 26.6
Gary Bell 6041 1062.77 1046.35 101.6 37.0
Jim Kaat 13638 2399.28 2362.22 101.6 85.9
Catfish Hunter 10176 1790.22 1763.00 101.5 64.0
Rheal Cormier 3292 579.15 570.51 101.5 24.2
Mike Scott 6182 1087.57 1071.79 101.5 46.7
Ken Brett 4613 811.55 800.11 101.4 30.1
Craig Lefferts 3396 597.44 589.91 101.3 21.0
Mike Norris 3394 597.09 589.74 101.2 24.3
Dave Giusti 4944 869.78 859.19 101.2 29.0
Mike Hampton 5555 977.27 965.48 101.2 37.6
Dock Ellis 6303 1108.86 1095.52 101.2 40.7
Matt Young 3610 635.09 627.59 101.2 22.1
John Burkett 7474 1314.87 1299.52 101.2 43.6
Steve Stone 5393 948.77 938.45 101.1 30.8
Lee Stange 3654 642.83 635.85 101.1 21.0
Oil Can Boyd 4204 739.59 731.67 101.1 25.2
Skip Lockwood 3693 649.69 643.28 101.0 24.9
Hideo Nomo 5349 941.03 932.56 100.9 35.9
Mike Moore 8541 1502.58 1489.11 100.9 56.9
Jim Rooker 5424 954.22 945.87 100.9 32.3
Bruce Kison 5424 954.22 946.33 100.8 30.5
Moose Haas 4937 868.55 861.95 100.8 29.2
Woody Williams 5004 880.33 873.84 100.7 31.5
Rick Wise 9412 1655.81 1643.89 100.7 53.5
Denny McLain 5645 993.10 985.97 100.7 39.0
Charlie Leibrandt 6913 1216.18 1208.18 100.7 42.0
Moe Drabowsky 4655 818.94 813.68 100.6 35.1
Dennis Leonard 6505 1144.40 1137.35 100.6 38.1
Mike Caldwell 7217 1269.66 1262.41 100.6 45.5
Paul Foytack 3575 628.94 626.42 100.4 20.0
Chris Hammond 3279 576.86 574.69 100.4 22.3
Joe Nuxhall 4352 765.63 762.75 100.4 24.9
Joey Jay 4444 781.81 779.56 100.3 26.9
Claude Osteen 10267 1806.23 1801.20 100.3 56.2
Darold Knowles 3243 570.53 569.09 100.3 21.0
Dick Ellsworth 6418 1129.09 1126.62 100.2 38.9
Jose Mesa 4047 711.97 710.91 100.1 32.5
Shawn Estes 4282 753.31 752.47 100.1 26.1
Denny Neagle 5324 936.63 935.76 100.1 33.3
Bud Black 6113 1075.44 1074.83 100.1 34.5
Julian Tavarez 3082 542.20 542.21 100.0 20.9
Gene Conley 3190 561.20 561.99 99.9 18.7
Pedro Ramos 6269 1102.88 1105.06 99.8 38.1
Dave Roberts 6270 1103.06 1105.36 99.8 39.2
Larry Christenson 4194 737.83 739.54 99.8 23.7
Mike Flanagan 8177 1438.55 1442.48 99.7 45.8
Ed Figueroa 3886 683.65 685.66 99.7 21.8
Miguel Batista 3122 549.24 551.04 99.7 17.9
Bob Purkey 5646 993.28 996.78 99.6 30.7
Robin Roberts 6207 1091.97 1096.40 99.6 35.9
Melido Perez 4034 709.69 712.62 99.6 24.9
Bob Tewksbury 5361 943.14 947.13 99.6 28.2
Rick Helling 3849 677.14 680.21 99.5 25.4
Reggie Cleveland 5472 962.67 967.08 99.5 29.6
Tom Bradley 3093 544.14 546.90 99.5 19.0
Bobby Bolin 4753 836.18 840.45 99.5 29.3
Lew Burdette 5874 1033.39 1039.25 99.4 32.2
Gary Peters 6185 1088.10 1094.57 99.4 37.3
Bob Moose 3792 667.11 671.13 99.4 22.5
Dave McNally 8056 1417.26 1427.97 99.3 42.6
Mike Morgan 7877 1385.77 1396.77 99.2 47.6
Curt Simmons 5399 949.82 957.57 99.2 32.3
Dan Petry 6188 1088.63 1097.67 99.2 35.4
Pat Dobson 6317 1111.32 1120.88 99.1 32.5
Darryl Kile 5890 1036.20 1045.61 99.1 31.2
Vern Law 5203 915.34 923.72 99.1 29.5
Steve Trout 4541 798.88 806.51 99.1 29.4
Rick Rhoden 7758 1364.83 1378.77 99.0 40.3
Tom Underwood 4731 832.31 840.96 99.0 25.4
Grant Jackson 4032 709.33 717.20 98.9 25.8
Dick Tidrow 5213 917.10 927.44 98.9 29.3
Juan Berenguer 3593 632.10 640.30 98.7 22.0
Dick Drago 5607 986.42 999.36 98.7 27.9
Andy Ashby 4805 845.32 857.21 98.6 29.3
Scott Sanderson 7687 1352.34 1372.16 98.6 41.2
Scott McGregor 6344 1116.07 1132.73 98.5 31.1
Dave Burba 4686 824.39 837.43 98.4 23.8
Pedro Astacio 5644 992.93 1009.39 98.4 29.3
Charlie Hough 11411 2007.49 2043.54 98.2 59.2
Esteban Loaiza 5278 928.54 946.32 98.1 29.2
Jim Merritt 4456 783.93 799.16 98.1 22.4
Doc Medich 5957 1047.99 1068.37 98.1 26.8
Steve Hargan 4927 866.79 883.73 98.1 34.2
Danny Jackson 6220 1094.26 1115.75 98.1 35.6
Sidney Ponson 3695 650.05 662.88 98.1 19.1
Dave Wickersham 3384 595.33 607.30 98.0 16.6
Russ Ortiz 3716 653.74 666.90 98.0 21.1
Cal McLish 3843 676.08 689.72 98.0 20.2
John Montefusco 4958 872.24 891.19 97.9 23.4
Vern Ruhle 4231 744.34 760.56 97.9 21.1
Ken Hill 5493 966.36 987.43 97.9 30.7
Dustin Hermanson 3167 557.16 569.47 97.8 18.8
Todd Stottlemyre 6278 1104.46 1129.02 97.8 29.0
Joe Magrane 3289 578.62 591.55 97.8 16.6
Orlando Pena 3624 637.56 652.11 97.8 19.3
John Buzhardt 4382 770.91 788.66 97.7 23.4
Pete Vuckovich 4318 759.65 777.15 97.7 20.5
Ron Darling 7097 1248.55 1277.47 97.7 32.6
Dave Stewart 7840 1379.26 1411.37 97.7 37.7
Shane Rawley 5631 990.64 1014.26 97.7 28.5
Al Fitzmorris 3797 667.99 684.23 97.6 18.3
Rick Sutcliffe 8136 1431.33 1466.22 97.6 44.7
Woodie Fryman 7175 1262.27 1293.58 97.6 38.7
Livan Hernandez 5203 915.34 938.81 97.5 25.5
Lary Sorensen 5203 915.34 939.30 97.4 24.0
Bill Lee 5876 1033.74 1060.99 97.4 27.7
Pete Harnisch 5267 926.60 951.46 97.4 26.8
Paul Quantrill 3565 627.18 644.04 97.4 22.2
Joe Coleman 7723 1358.68 1396.30 97.3 40.0
Dick Ruthven 6386 1123.46 1154.69 97.3 28.1
Denny Lemaster 5048 888.07 912.99 97.3 21.3
Wally Bunker 3227 567.71 584.18 97.2 16.4
Charles Hudson 3054 537.28 552.98 97.2 14.7
Ken Holtzman 8560 1505.93 1549.97 97.2 40.8
Steve Trachsel 5822 1024.24 1054.41 97.1 24.1
Bob Turley 3011 529.71 545.38 97.1 14.3
Jim Bouton 3697 650.40 669.88 97.1 18.1
Steve Renko 7484 1316.63 1356.77 97.0 32.3
Roger Craig 3772 663.59 683.88 97.0 21.1
Joaquin Andujar 6381 1122.58 1157.25 97.0 28.6
Pedro Borbon 3027 532.53 549.54 96.9 14.5
Cal Eldred 3779 664.82 686.13 96.9 17.7
Rick Langford 4489 789.73 815.61 96.8 22.2
Hal Brown 3201 563.14 581.63 96.8 14.3
Jim Clancy 7532 1325.07 1371.52 96.6 34.8
Ernie Broglio 3959 696.49 721.24 96.6 21.7
Jack Billingham 6605 1161.99 1204.17 96.5 30.4
Ted Power 3472 610.81 633.03 96.5 17.7
Terry Mulholland 7225 1271.06 1318.44 96.4 35.6
Mudcat Grant 7297 1283.73 1331.75 96.4 32.2
Jim Colborn 4791 842.86 874.90 96.3 20.2
Mark Clark 3514 618.20 641.78 96.3 16.3
Floyd Bannister 7154 1258.57 1308.15 96.2 30.4
Diego Segui 5391 948.42 985.96 96.2 22.5
Kent Mercker 3466 609.76 633.90 96.2 18.2
Mike McCormick 7060 1242.04 1291.29 96.2 29.5
Fred Norman 5766 1014.39 1054.80 96.2 25.6
Bob McClure 3455 607.82 632.09 96.2 18.4
Don Cardwell 6065 1066.99 1109.67 96.2 25.1
Juan Pizarro 5960 1048.52 1090.75 96.1 29.9
Larry Gura 6051 1064.53 1109.20 96.0 27.0
Frank Castillo 4766 838.46 873.66 96.0 21.9
Kirk Rueter 5146 905.31 944.36 95.9 23.5
Wayne Garland 3112 547.48 571.14 95.9 15.9
Doyle Alexander 9989 1757.32 1835.31 95.8 45.6
Carl Morton 4862 855.35 893.34 95.7 19.9
Paul Lindblad 3602 633.69 661.91 95.7 18.2
Walt Terrell 5927 1042.71 1089.64 95.7 23.0
Bob Shaw 5230 920.09 961.72 95.7 24.0
Jim Lonborg 7448 1310.30 1369.96 95.6 30.2
Bob Bruce 3384 595.33 622.53 95.6 14.7
Bill Gullickson 7644 1344.78 1406.74 95.6 30.1
Clyde Wright 5142 904.61 946.77 95.5 20.5
Bob Shirley 4267 750.68 785.68 95.5 17.2
Pat Zachry 3535 621.90 651.45 95.5 14.9
Omar Olivares 4158 731.50 767.10 95.4 17.5
Aaron Sele 5354 941.91 987.90 95.3 23.0
Stan Bahnsen 7606 1338.09 1403.94 95.3 29.9
Bruce Ruffin 3844 676.26 709.88 95.3 25.0
Joe Gibbon 3224 567.19 595.77 95.2 14.9
Bill Wegman 4424 778.30 817.60 95.2 18.0
Mark Leiter 3558 625.94 658.21 95.1 14.1
Mike Krukow 6620 1164.63 1225.09 95.1 23.3
Rick Mahler 5859 1030.75 1085.46 95.0 20.4
Mark Davis 3460 608.70 641.31 94.9 23.4
Don Robinson 5852 1029.52 1085.39 94.9 23.0
Dick Bosman 4759 837.23 882.83 94.8 22.7
Craig Swan 3692 649.52 685.38 94.8 19.1
Bob Buhl 5511 969.53 1023.06 94.8 21.4
Tom Browning 5770 1015.09 1071.63 94.7 20.1
Paul Splittorff 7589 1335.10 1410.36 94.7 27.9
Tom Hume 3258 573.17 605.71 94.6 16.0
Eddie Fisher 4546 799.76 846.26 94.5 19.8
Bobby Witt 6901 1214.06 1284.76 94.5 24.5
Phil Regan 4128 726.22 768.72 94.5 19.7
Ed Whitson 6693 1177.47 1247.11 94.4 26.2
Dan Schatzeder 3924 690.33 731.47 94.4 17.5
Mike Torrez 9075 1596.53 1691.75 94.4 34.5
Nelson Briles 6311 1110.27 1176.97 94.3 22.6
Ray Washburn 3592 631.93 670.82 94.2 14.4
Andy Hassler 3361 591.29 627.73 94.2 15.6
Kevin Gross 7516 1322.26 1404.61 94.1 24.0
Milt Wilcox 5964 1049.22 1114.69 94.1 20.8
Ron Kline 4557 801.69 851.95 94.1 17.5
Bobby Jones 4354 765.98 814.30 94.1 14.5
Larry McWilliams 4681 823.51 876.92 93.9 20.0
Jim Deshaies 4569 803.81 856.76 93.8 17.7
Glen Hobbie 3752 660.07 704.45 93.7 12.7
Danny Cox 3829 673.62 719.26 93.7 17.2
Pete Schourek 3113 547.66 585.06 93.6 13.3
Doug Rau 3734 656.91 702.02 93.6 12.2
Mark Gardner 4856 854.30 913.16 93.6 14.2
Dan Spillner 4514 794.13 849.44 93.5 17.5
Al Jackson 4169 733.44 785.27 93.4 13.9
Eric Show 4915 864.68 926.76 93.3 16.4
Barry Latman 3631 638.79 684.73 93.3 16.1
Eric Milton 3520 619.26 664.58 93.2 11.2
Bob Knepper 8115 1427.64 1534.86 93.0 29.4
Doug Bird 3615 635.97 683.86 93.0 13.1
Steve Sparks 3505 616.62 663.50 92.9 10.9
Brett Tomko 3753 660.25 711.76 92.8 11.6
Steve McCatty 3581 629.99 679.37 92.7 12.2
Jim McGlothlin 3849 677.14 730.70 92.7 14.5
Steve Blass 4587 806.97 871.21 92.6 16.5
Mike Bielecki 3715 653.56 705.89 92.6 11.1
Ray Sadecki 7503 1319.97 1426.63 92.5 24.3
Gene Nelson 3208 564.37 610.03 92.5 13.0
Clay Kirby 4566 803.28 868.59 92.5 14.7
George Brunet 4220 742.41 803.60 92.4 16.7
Ster. Hitchcock 3252 572.11 619.58 92.3 8.4
Bob Forsch 8268 1454.56 1577.62 92.2 24.1
Jason Bere 3138 552.06 599.78 92.0 11.2
Brian Anderson 4233 744.69 810.06 91.9 10.4
Joe Niekro 10695 1881.53 2048.44 91.9 29.6
Glendon Rusch 3894 685.06 746.79 91.7 10.8
Bill Stoneman 3673 646.18 704.45 91.7 13.1
Sammy Ellis 3037 534.29 582.90 91.7 11.6
Dennis Rasmussen 4358 766.69 837.37 91.6 18.2
Richard Dotson 5529 972.69 1062.65 91.5 15.3
Tom Murphy 4264 750.15 819.83 91.5 14.1
Jason Johnson 3288 578.44 632.97 91.4 6.9
Tom Phoebus 3040 534.81 585.24 91.4 7.8
Tim Leary 4480 788.15 862.51 91.4 11.5
Mike LaCoss 5156 907.07 994.25 91.2 13.5
Lynn McGlothen 4499 791.49 867.98 91.2 13.3
Tony Cloninger 5005 880.51 966.02 91.1 12.6
Ramon Ortiz 3040 534.81 589.43 90.7 5.7
Darren Oliver 3677 646.88 713.27 90.7 9.5
Pat Rapp 3697 650.40 717.18 90.7 7.9
Willie Blair 3411 600.08 662.53 90.6 8.2
Bob Walk 4996 878.93 970.50 90.6 10.1
Curt Young 3304 581.26 642.95 90.4 9.0
Neal Heaton 4509 793.25 878.34 90.3 9.7
John Curtis 4929 867.14 960.59 90.3 10.0
Chuck Dobson 3763 662.01 733.41 90.3 7.3
Matt Keough 3589 631.40 700.01 90.2 8.8
Jaime Navarro 5664 996.44 1105.40 90.1 14.2
Jim Nash 3310 582.31 646.25 90.1 8.5
James Baldwin 3354 590.06 655.00 90.1 8.4
Bl. Mo. Odom 4504 792.37 882.02 89.8 13.3
Pete Smith 3091 543.79 605.46 89.8 6.2
Alan Foster 3054 537.28 598.79 89.7 8.8
Jim Slaton 7998 1407.06 1571.00 89.6 13.1
Jack Fisher 5892 1036.56 1158.35 89.5 10.6
Bill Krueger 3608 634.74 709.92 89.4 7.1
Eric Rasmussen 3009 529.36 592.52 89.3 6.7
Mike Smithson 4070 716.02 802.27 89.2 5.3
Glenn Abbott 3841 675.73 757.22 89.2 7.6
Bill Travers 3368 592.52 664.16 89.2 5.8
Jeff Suppan 4336 762.81 855.31 89.2 6.4
Wayne Twitchell 3153 554.69 623.47 89.0 9.9
Ray Burris 6579 1157.42 1305.81 88.6 11.7
Tom Griffin 4517 794.66 899.22 88.4 9.6
Jamey Wright 3674 646.35 733.92 88.1 4.2
George Stone 3018 530.94 606.17 87.6 4.6
Ross Grimsley 5928 1042.89 1194.10 87.3 6.9
Lew Krausse 3835 674.68 776.95 86.8 3.8
Jimmy Haynes 3151 554.34 639.28 86.7 5.5
Andy Hawkins 4625 813.66 938.39 86.7 3.2
Dave LaPoint 4419 777.42 896.82 86.7 5.6
Ricky Bones 3659 643.71 743.54 86.6 4.0
Jose Lima 3900 686.11 797.42 86.0 3.5
Jesse Jefferson 3287 578.27 672.13 86.0 3.5
Dave Mlicki 3079 541.68 646.35 83.8 -0.4
Pete Falcone 4256 748.74 899.40 83.2 0.0
Randy Lerch 3248 571.41 736.10 77.6 -8.1


I'd like to understand these numbers a little better.

a) Is DNRA+ similar to adjusted ERA+, but with defense factored in? If it is, then how many points of DNRA+ would you expect a pitcher to lose for every 1000 outs more?

b) Is the marker number similar to something like WAR or WARP, where you try to assign a value for each season, and then just add them all together to get the pitcher's career total?

c) Could I please get a brief explanation of XRA and NRA?


Thanks Matt.

[SABR Matt]

First XRA and NRA:

XRA is how many runs a pitcher should give up if he's exactly average in a league where 4.75 runs score per game. Simply multiply 4.75 by the pitcher's number of defense independent outs and divide by 27. DNRA is a metric designed to track actual runs...not earned runs so everything is scaled to actual runs and the errors that go on behind a pitcher are assumed to occur at the league average rate per batted ball in play (by default, this assumption occurs when you use baseRuns to evaluate the expected run-value of each possible event and then assume that anything not in the pitcher's control happens at the league average rate plus the rate at which he allows those things to happen above that which occurs when every other pitcher on his team is on the mound).

NRA is the pitcher's defense neutral runs allowed adjusted to that same 4.75 R/G context. DNRA calculates everything calibrated to the league average run scoring rate, so the original baseRuns value must be multiplied by Lg Runs per Game / 4.75 (the odds ratio method) to obtain a common run scoring context for all pitchers.

DNRA+ is 100*(XRA / NRA), yielding an ERA+ like metric where 100 is average and higher is better, but with defense factored out (along with the unfair weighting of park factors that goes on in traditional ERA+).

The Marker number is an estimate of the player's worthiness to be in the HOF. It's found with the following steps:

Calculate the pitcher's marginal run value in four steps:

1) Pitcher's Winning Percentage:

----> PWP = 1 - (NRA^X / (NRA^X + XRA^X)); X = ((NRA+XRA) / (DIOuts / 27))^0.285

2) Find the offense-equivalent run margin:

----> Margin = (.35 / .65)^(1/X) * XRA (this forces the margin to be defined using pythagorean logic at a .350 winning percentage)

3) Find the pitcher's marginal run value:

----> Offense Equivalent Runs (OER) = (PWP / (1 - PWP))^(1/X) * XRA
----> Marginal Runs (MR) = OER - Margin

4) Convert the pitcher's marginal runs to wins and the wins to a marker score:

----> W = MR / 10
----> Marker = W*2 - Average W
----> Average W = (XRA - Margin) / 10

The theory being that a player's HOF worthiness can be estimated in short as his wins in each season plus his wins above average in each season (giving him credit for being well above average on top of credit for the value he produced).

If you've got any other questions or concerns, feel free to let me know and hopefully I can be of help.

[SABR Matt]

Just noticed I missed a question...

I have debated some with STLCards2 and brett in the past about how best to deal with the reality that the more someone pitches, the harder it is to be dominant (rise far above average)...I'm not sure there's a clear answer to that question...at least I don't think we've found any kind of consensus best idea yet. I, personally, believe the best place to start would be to use Tom Tango's favorite approach and regress everything to the mean once we've picked a common playing time length to which everyone should be compared. But that's risky and may make incorrect assumptions so I'm not totally comfortable doing it as yet.

[Matthew C.]

Matt - to clarify, DNRA+ has seen a lot of changes over the years. The current edition...

Does account for variances vs. team regarding XBH prevention and pitcher defense.

It does not account for situational splits or pitcher offense.

Is this correct?

And when will 1999, 2005-2009 be included? Isn't 1999 batted ball data available yet?

[curveball]

Some of your numbers had me wondering, so I have a few questions that would probably clear things up better for me

a) How does Nolan Ryan go from a career era+ of 111, and overtake Tom Seaver, who had a 127? Ryan goes to 121.7, and Seaver drops to 120.6. I didn't go through the whole list, but I definitely noticed that Ryan jumped 10 points. The same goes for Pettitte, as he jumped about 9 points, so I am assuming that they are 2 of the biggest jumpers. What was it about Ryan that he got shortchanged so much by era+ over his career?

b) You have Marichal and Gibson in a virtual dead heat. Something like WAR has Gibson ahead 85.5 to 64.1. PRAA has Gibson ahead 349 to 221. I believe both WAR and PRAA use defensive considerations. Who was better in your opinion based on your findings?

c) You probably would have answered this question by answering questions A and B. I noticed that Phil Niekro took a serious hit in your findings. He had a 96.8 WAR to Ryan's 84.8, and a 359 PRAA to Ryan's 222. Based on your research you would seem to have this group of pitchers rated kind of in this order then?

1 Nolan Ryan
2 Tom Seaver
3 Gaylord Perry
4 Bert Blyeven
5 Steve Carlton
t6 Bob Gibson
t6 Juan Marichal
8 Fergie Jenkins
9 Don Sutton
10 Jim Palmer
11 Tommy John
12 Phil Niekro


Thanks again Matt.

[SABR Matt]

Matt - to clarify, DNRA+ has seen a lot of changes over the years. The current edition...

Does account for variances vs. team regarding XBH prevention and pitcher defense.

It does not account for situational splits or pitcher offense.

Is this correct?

And when will 1999, 2005-2009 be included? Isn't 1999 batted ball data available yet?

DNRA hasn't actually seen a lot of changes, STLCard2...the only things I've changed since the first draft hit fever in 2006 are:

I switched from linear weights to baseRuns after Tom Tango persuaded me that the pitcher's own run environment was more accurate than the average one.

I began considering pitcher-fielded balls as defense independent pitching successes.

I corrected my calculation of career DNRA+ to account for changing run environments and put everyone on the same scale.

I fixed a small bug in my calculation of the pythag exponent.

The system currently accounts for the following:

- The pitcher's ability to prevent in play hits of all types relative to his team defenses
- The pitcher's DIPS skill
- The pitcher's ability to check the running game
- The pitcher's ability to field his own position
- The pitcher's park-adjusted HR rate

No accounting for the strength of opponent faced or situational variability are included though I'd like to address both.

1999 data is available and I've already downloaded it. 1952-1956 are also available and I've stored them as I've done with 2006-2009. I just haven't had time to recalculate DNRA+ for the new database. I'm hoping to get to that this winter as I'm tired of having to hand-calculate a DNRA when people ask me about a current pitcher's results. :)

[SABR Matt]

Some of your numbers had me wondering, so I have a few questions that would probably clear things up better for me

a) How does Nolan Ryan go from a career era+ of 111, and overtake Tom Seaver, who had a 127? Ryan goes to 121.7, and Seaver drops to 120.6. I didn't go through the whole list, but I definitely noticed that Ryan jumped 10 points. The same goes for Pettitte, as he jumped about 9 points, so I am assuming that they are 2 of the biggest jumpers. What was it about Ryan that he got shortchanged so much by era+ over his career?

b) You have Marichal and Gibson in a virtual dead heat. Something like WAR has Gibson ahead 85.5 to 64.1. PRAA has Gibson ahead 349 to 221. I believe both WAR and PRAA use defensive considerations. Who was better in your opinion based on your findings?

c) You probably would have answered this question by answering questions A and B. I noticed that Phil Niekro took a serious hit in your findings. He had a 96.8 WAR to Ryan's 84.8, and a 359 PRAA to Ryan's 222. Based on your research you would seem to have this group of pitchers rated kind of in this order then?

1 Nolan Ryan
2 Tom Seaver
3 Gaylord Perry
4 Bert Blyeven
5 Steve Carlton
t6 Bob Gibson
t6 Juan Marichal
8 Fergie Jenkins
9 Don Sutton
10 Jim Palmer
11 Tommy John
12 Phil Niekro


Thanks again Matt.

:)

I smile because we've discussed Nolan Ryan so much over the last year that I'm starting to hate that guy...LOL The biggest moves in the system that I can find are Mussina (+11), Ryan (+10) and Pettitte (+9)...they have a few things in common:

1) They pitched in predominantly pitcher's parks. Park factor analysis at baseball-reference.com is rather rough...the same park run factor is applied to everyone's ERA+, but the balls in play you allow, the less the park affects your ability to prevent runs. So the most defense independent pitchers don't get the kind of park nudge that contact pitchers get for being at a pitcher's park like YS, the Astrodome etc.

2) They pitched in front of lousy defenses during their best seasons. Mussina and Pettitte in particular has some seriously awful team defenses behind them.

3) There was some other critical factor for each of them...for Ryan, his bullpens were among the worst you'll ever find for someone who lasted as long as he did. Many of his runners (left after he was pulled from games) scored because the guy who came in to replace him sucked. For Pettitte, he was good at a lot of little things - baserunning prevention, controlling his wildness on things like WP, HBP etc and inducing the double play - Pettitte also didn't move as much as you may think...his ERA+ is CURRENTLY 116, but that includes 1999 (101), 2006 (106), 2007 (111), 2008 (98), and 2009 (103). If you take those years out of his ERA, his ERA+ rises to 120.

As for why Neikro takes such a massive hit, he may be the kind of pitcher that benefits most from situational split analysis (while Ryan might suffer a bit from the same). Neikro, in a DIPS sense is rather poor...his K/BB is a bit low and his HR rate is way WAY too high. He allowed 298 solo HRs and 178 HRs with men on base...I don't have a feel for whether that's the normal percentage. Everything else about his bases empty vs. men on base split looks identical if you take out the intentional walks, which DNRA charges less for than regular walks. So...I am currently doubting that Neikro benefits much form great situational pitching. I suspect in his case, the main thing is that he pitched in hitter's parks and was defense independent enough to get too much credit for that at b-ref.

[SABR Matt]

BTW, part of your concerns may be mollified by a misunderstanding about the scale of the Marker variable. You're comparing WAR to the marker...the marker isn't a measure of career wins...the marker is a measure of greatness...pitchers with a shorter career who produce identical WAR values will have a higher marker score as that implies they were farther above average during their career so their marker to WAR exchange rate will be higher.

In other words...if you're a better pitcher in your prime, you get more credit for that prime.

I do this rather than show straight WAR because I think it is problematic to use an ERA+-like stat and innings pitched and assume that the shape of the career that produced that pair of numbers doesn't matter when judging a pitcher's relative greatness. I'll see people compare a short-careered pitcher like Dizzy Dean to a long-careered pitcher like Phil Neikro and they never seem to properly account for the fact that being way above average has non-linear value to a franchise.

[brett]

DNRA hasn't actually seen a lot of changes, STLCard2...the only things I've changed since the first draft hit fever in 2006 are:

I switched from linear weights to baseRuns after Tom Tango persuaded me that the pitcher's own run environment was more accurate than the average one.

I began considering pitcher-fielded balls as defense independent pitching successes.

I corrected my calculation of career DNRA+ to account for changing run environments and put everyone on the same scale.

I fixed a small bug in my calculation of the pythag exponent.

The system currently accounts for the following:

- The pitcher's ability to prevent in play hits of all types relative to his team defenses
- The pitcher's DIPS skill
- The pitcher's ability to check the running game
- The pitcher's ability to field his own position
- The pitcher's park-adjusted HR rate

No accounting for the strength of opponent faced or situational variability are included though I'd like to address both.

1999 data is available and I've already downloaded it. 1952-1956 are also available and I've stored them as I've done with 2006-2009. I just haven't had time to recalculate DNRA+ for the new database. I'm hoping to get to that this winter as I'm tired of having to hand-calculate a DNRA when people ask me about a current pitcher's results. :)

As for the pythagorian calculation, you wrote that you convert pitcher runs saved into runs added (I think). Isn't a run saved worth MORE than a run added because there are a limited number of runs that can be saved?

And just from the pythag calculation isn't a run saved worth a little more than a run produced?

[brett]

:)

I smile because we've discussed Nolan Ryan so much over the last year that I'm starting to hate that guy...LOL The biggest moves in the system that I can find are Mussina (+11), Ryan (+10) and Pettitte (+9)...they have a few things in common:

1) They pitched in predominantly pitcher's parks. Park factor analysis at baseball-reference.com is rather rough...the same park run factor is applied to everyone's ERA+, but the balls in play you allow, the less the park affects your ability to prevent runs. So the most defense independent pitchers don't get the kind of park nudge that contact pitchers get for being at a pitcher's park like YS, the Astrodome etc.

2) They pitched in front of lousy defenses during their best seasons. Mussina and Pettitte in particular has some seriously awful team defenses behind them.

3) There was some other critical factor for each of them...for Ryan, his bullpens were among the worst you'll ever find for someone who lasted as long as he did. Many of his runners (left after he was pulled from games) scored because the guy who came in to replace him sucked. For Pettitte, he was good at a lot of little things - baserunning prevention, controlling his wildness on things like WP, HBP etc and inducing the double play - Pettitte also didn't move as much as you may think...his ERA+ is CURRENTLY 116, but that includes 1999 (101), 2006 (106), 2007 (111), 2008 (98), and 2009 (103). If you take those years out of his ERA, his ERA+ rises to 120.

As for why Neikro takes such a massive hit, he may be the kind of pitcher that benefits most from situational split analysis (while Ryan might suffer a bit from the same). Neikro, in a DIPS sense is rather poor...his K/BB is a bit low and his HR rate is way WAY too high. He allowed 298 solo HRs and 178 HRs with men on base...I don't have a feel for whether that's the normal percentage. Everything else about his bases empty vs. men on base split looks identical if you take out the intentional walks, which DNRA charges less for than regular walks. So...I am currently doubting that Neikro benefits much form great situational pitching. I suspect in his case, the main thing is that he pitched in hitter's parks and was defense independent enough to get too much credit for that at b-ref.

Wasn't there an argument that knuckeballers have more control over BABIP?

[SABR Matt]

As for the pythagorian calculation, you wrote that you convert pitcher runs saved into runs added (I think). Isn't a run saved worth MORE than a run added because there are a limited number of runs that can be saved?

And just from the pythag calculation isn't a run saved worth a little more than a run produced?

Yes.

The way I convert runs allowed into an offense-equivalent run scored is through the pythag calculation. I take the pitcher's winning percentage (which is also a pythag calculation and which will go higher faster (per run saved) than it would go lower (per run allowed above average) for the first big chunk of possible runs allowed because W% approaches 1 faster than it approaches zero...and I use it to recalculate runs in the form of offense instead of defense allowed. So if you give up half the runs allowed that we expect, your pythag W% will be somewhere near .800 and you'll be credited with twice the runs scored of an average pitcher.

[SABR Matt]

Wasn't there an argument that knuckeballers have more control over BABIP?

Yes...but DNRA will SEE that...it should reveal whether Neikro was preventing more base hits on balls in play than his team.

[Matthew C.]

Matt - what would you like to see in terms of DNRA marker for HOF worthiness? Around 100?


From a few posts ago, reagarding Niekro: "He allowed 298 solo HRs and 178 HRs with men on base...I don't have a feel for whether that's the normal percentage"


As far as a normal %, I have no idea, but...

Niekro is 1 HR/52 PA's with men on and 1 HR/44 PA's with nobody on.

Doesn't seem too far out of the normal range.

Compare that to the guy most famed for preventing HR's with men on compared to nobody on:

Tom Glavine - 1 HR/44 PA's with nobody on, 1 HR/68 PA's with one on, 1 HR/75 PA's with 2 on, and 1 HR/ 214 PA's with bases loaded. :eek:

Now those are meaningful splits. Where guys like Glavine get hurt concerning linear weights (or even BaseRuns) is that they assume an even distribution of HR's allowed. Glavine and a few others gave up a whole lot fewer runs on their homeruns than the average pitcher would have given the same number of HR's allowed. Niekro doesn't look like one of those guys. I know a few years back when I was researching sit. splits there was one other guy who had anything that looked even remotely similar to this, but I can't remember who it was. I know Palmer is famous for never allowing a GS, but I don't remember if his whole men/on resume looked like that or not.

Also, Niekro's BABIP dropped from .275 with BE to .266 with RISP, for what its worth.

I'm going to check out Maddux to see what his line looks like.

[Matthew C.]

Maddux was actually a little worse with men on than with nobody on (1 HR/54 PA's compared to 1 HR/60 PA's). Palmer was about the same both ways - even with his 0 HR/ 213 PA's w/ bases loaded.

[brett]

Maddux was actually a little worse with men on than with nobody on (1 HR/54 PA's compared to 1 HR/60 PA's). Palmer was about the same both ways - even with his 0 HR/ 213 PA's w/ bases loaded.

Maddux actually has the worst OPS+ allowed split with runners on base that I have found: 115 with runners on base and 91 with the bases empty.

Ryan was 113 and 91

Most guys with fairly long careers are in the 103-107 range for OPS+ allowed with runners on base.

[SABR Matt]

I could certainly apply a situational DIPS adjustment without fear of confounding the data. Since the pitcher is always in control of the DIPS events so there's no risk of a situational split being defense-generated luck...it would be within the original spirit of DNRA to record the things the pitcher actually did to influence the outcome of the game if I broke all of the DIPS stats up by base/out state and gave a situational run-expectancy-added value to those, rather than a flat linear weight.

I would bet that this captures about 95% of the total (real) ability a pitcher has to influence run-scoring through situational pitching, as a matter of fact. Most, if not all, of the examples I've seen of pitchers who are significantly difference with men on base express themselves along the DIPS line. Glavine prevents HRs and gives up more walks (and gets more Ks) with men on base, for example. Ryan throws a lot of wild pitches, hits batters more with men on base (I just noticed that looking at his split), and walks more guys without a corresponding increase in K rate.

I believe that is about as far as I want to take situational analysis in DNRA...doing anything else will become too speculative, especially from year to year.

[brett]

I could certainly apply a situational DIPS adjustment without fear of confounding the data. Since the pitcher is always in control of the DIPS events so there's no risk of a situational split being defense-generated luck...it would be within the original spirit of DNRA to record the things the pitcher actually did to influence the outcome of the game if I broke all of the DIPS stats up by base/out state and gave a situational run-expectancy-added value to those, rather than a flat linear weight.

I would bet that this captures about 95% of the total (real) ability a pitcher has to influence run-scoring through situational pitching, as a matter of fact. Most, if not all, of the examples I've seen of pitchers who are significantly difference with men on base express themselves along the DIPS line. Glavine prevents HRs and gives up more walks (and gets more Ks) with men on base, for example. Ryan throws a lot of wild pitches, hits batters more with men on base (I just noticed that looking at his split), and walks more guys without a corresponding increase in K rate.

I believe that is about as far as I want to take situational analysis in DNRA...doing anything else will become too speculative, especially from year to year.

I do think anything else would be too speculative.

Ryan, btw had a much lower K rate with runners on base (22.6% versus 27.3%). That a) seems to explain a lot and b) could very well be due to pitching from the stretch.

In comparison Seaver was 18.6% compared to 18.9%.

[SABR Matt]

That also slightly (very slightly) reduces the linear weight value of the average Ryan punchout.

Reduced K rate, increased walk rate (not all of which was IBB) and increased HR rate, with more HBP and of course the WP too (all of which happen with men on base except for the occasional K/WP).

I am confident you will see any significant style change that accompanies the base/out state splits directly revealed in the DIPS statistics and any other fluctuations will either be too small to worry about or difficult to pin down who's to blame. Afterall, if there's an unusual increase in singles or doubles or whatnot when the pitcher has guys on base, it might be due to the pitcher and fielders not being on the same page that particular year with regard to how to position to defend behind a pitcher's preferred style with runners on. Or something else that's not 100% the fault of the pitcher.

[brett]

If you account for that, I'd never look at ERA+ again.

By the way, I was wondering if relievers tend to face a slightly higher level of hitter, or tougher handed splits that starters. And how do we account, for example for someone like Koufax facing a much higher than normal percentage of left handed hitters versus the average of his time? Teams even threw inferior lefty hitters against him.

[Matthew C.]

That also slightly (very slightly) reduces the linear weight value of the average Ryan punchout.

Reduced K rate, increased walk rate (not all of which was IBB) and increased HR rate, with more HBP and of course the WP too (all of which happen with men on base except for the occasional K/WP).

I am confident you will see any significant style change that accompanies the base/out state splits directly revealed in the DIPS statistics and any other fluctuations will either be too small to worry about or difficult to pin down who's to blame. Afterall, if there's an unusual increase in singles or doubles or whatnot when the pitcher has guys on base, it might be due to the pitcher and fielders not being on the same page that particular year with regard to how to position to defend behind a pitcher's preferred style with runners on. Or something else that's not 100% the fault of the pitcher.

So your saying that Neikro's .007 BABIP drop with RISP compared to nobody on (even after a bazillion BF's) is probably not a reflection on his own ability, but probably more on defensive positioning, etc.? How much of a BABIP drop with how many BF's do we need to observe before it looks like as much "skill" as it does some type of defensive phenomenon? I am not saying you are wrong, just wondering at what point is their enough of a difference before you say "that has to be partially skill."

[SABR Matt]

.007 is a pretty darned tiny number. I know it's a lot of balls in play accumulated in those situations...and statistically, it would probably show up as significantly different in the sense that random chance wouldn't explain it...but significant and useful/relevant aren't the same thing. 0.007 BABIP points might (MIGHT) turn into a few runs saved over the course of a career...or...those BABIP points might be because hitters put the ball in the AIR more against Neikro anyway and with guys on base, the defense played back to cut off the double, thus turning doubles into sac flies, flyouts or HRs and not much in between...know what I mean? I think there is too much room for open speculation as to what would cause a tiny change like that for me to want to definitively include it in my metric.

I mean you could be right that Neikro had some BABIP skill with men on base...but...if he did...it was a pretty minor skill...

[brett]

.007 is a pretty darned tiny number. I know it's a lot of balls in play accumulated in those situations...and statistically, it would probably show up as significantly different in the sense that random chance wouldn't explain it...but significant and useful/relevant aren't the same thing. 0.007 BABIP points might (MIGHT) turn into a few runs saved over the course of a career...or...those BABIP points might be because hitters put the ball in the AIR more against Neikro anyway and with guys on base, the defense played back to cut off the double, thus turning doubles into sac flies, flyouts or HRs and not much in between...know what I mean? I think there is too much room for open speculation as to what would cause a tiny change like that for me to want to definitively include it in my metric.

I mean you could be right that Neikro had some BABIP skill with men on base...but...if he did...it was a pretty minor skill...

What about Niekro (and Quizenberry for example) creating GIDP?

Quiz was about 3%. Niekro 2%. Ryan 1.5%

[SABR Matt]

If you account for that, I'd never look at ERA+ again.

By the way, I was wondering if relievers tend to face a slightly higher level of hitter, or tougher handed splits that starters. And how do we account, for example for someone like Koufax facing a much higher than normal percentage of left handed hitters versus the average of his time? Teams even threw inferior lefty hitters against him.

Ideally, the next version of DNRA will attempt to quantify the quality of batter faced by each pitcher. The BABIP dataset allows us as "how did player X hit when not facing this particular pitcher?" (thus taking out the pitcher's influence on each batting line) and we can composite the strengths of all batters faced and turn it into an adjustment that would be applied to XRA.

I think ACE relievers do tend to face slightly better hitters overall, but that may be countered by the fact that they tend to face platoon-favored hitters too. Especially after about 1985 when the LOOGY became more popular.

[SABR Matt]

What about Niekro (and Quizenberry for example) creating GIDP?

DNRA accounts for that. DP rate (per runner at first base) is one of the components tracked by DNRA.

[brett]

Matt, I don't see this as a PCA question but I wondered where to find data on this question:

Do better pitchers tend "inspire" more run support-consistently.

The example that came up was Carlton in '72. He had a great pitching year of course, but it still took better than average run support (for his particular team) to get him to 27-10 and, in fact they scored something like 30% more for him.

I don't think that teams get inspired by aces, but in this case, where he was basically the only thing going for the team that year, it might have motivated the team a little to have him pitching on a particular night.

[SABR Matt]

I don't think a comprehensive study of the affects of "THE guy" on a team has ever been done. I personally believe that a belief system in a clubhouse is necessary for the players to perform up to their ability. Having just witnessed the 2008 Mariners go 61-101 and a team that wasn't all that much better on paper go 85-77 in 2009 and do so primarily by playing a lot harder day in and day out with the belief that they could hang with anyone...I think that what you're seeing with the 72 Phillies is a clubhouse with a losing culture (whose players were better than their performance) EXCEPT when Cartlton pitched and they really thought they could win.

And I think to see that affect in the data requires a special combination of circumstances (you need a clubhouse that is in bad shape AND an ace pitcher at the same time)...otherwise, the momentum you get from the ace will be the same as the momentum you get from your already outstanding clubhouse.

For example, Felix Hernandez put it all together this year and had a great season...he went 19-5 and the Mariners went something like 26-7 in his starts..but their offensive production CERTAINLY wasn't HIGHER in Hernandez starts...he actually got some of the worst run support on the team. It's not that he didn't have a strong positive influence on clubhouse chemistry and personal belief...it's that the Mariners in 2009 had that working for them already...with or without King Felix.

[leewileyfan]

Taking an historic look at pitchers from the standpoint of my evaluations, I divided the classes of starters and relievers as fairly as I could by dominant numbers in either role over the other. In a few cases, the selection may seem arbitrary, as with Rollie Fingers and Hoyt Wilhelm who served both functions. I put Fingers in the reliever set because his relief calls were more dominant; and putting him in that category expressed his "greatness" in the better light.

Wilhelm also had dominant relief calls; but, in his case, the sheer volume of games and innings skewed him value-wise [great in either function, but better than Mariano Rivera overall/???]. He is more fairly considered among the starters. I did not want to create a third list, hybrids, so here goes.

I. The Relievers

Each is presented with my adjusted ERA in an attempt to normalize pitchers across several generations. The ERA factor is then applied to career durations [years-innings] to produce a "value' or "greatness" rating.

Reliever..................Adj. ERA............Value #

Rivera......................1.56.................. 355.49
Percival....................2.22.................. 163.48
Hoffman...................2.12..................27 2.08
Dibble......................2.72.................. ..94.56
Sutter.....................2.80................... 197.04
Gossage...................3.07...................2 87.05
M. Jackson...............3.20...................157.1 0
Nen.........................2.98.................. .108.34
Franco.....................2.86................... 209.04
Quisenberry...............2.81.................... 95.91
Dotel........................3.14................. ...88.60
Fingers......................2.78................. .325.24
Worrell.....................3.15.................. .103.75
Wilks........................2.93................. ..158.84
Hall..........................2.90................ ...223.83
Murphy.....................3.31................... 138.14
Face........................4.02.................. . 74.09
Konstanty.................3.65................... 89.06
Lyle.........................3.04................. ...224.93
Marshall....................3.31.................. ..183.81
F. Rodriguez...............2.01...................143 .93
S. Miller....................3.09.................... 264.47
"B." Wagner...............1.65....................263.8 2
Myers.......................2.96.................. ..153.21
Foulke......................2.37.................. ...184.75
Montgomery...............2.88....................1 56.25
Reardon....................2.67................... ..230.08
Hrabosky...................3.06................... .115.44
McMahon..................3.00..................... 218.62
L. McDaniel................3.57....................22 0.13
Wetteland.................2.50.................... .170.41
Aker.........................3.56................. .... 78.13
H. Casey...................3.19.....................1 36.70
"A." Adams................3.46.......................64 .06
Beggs.......................2.59.................. ....147.27
Ferrick.......................3.27................ .......92.31
J. Heving....................4.07.................... ...49.12
Maltzberger................2.43................... .....67.57
Page.........................3.76................. .......66.06
Brantley.....................3.12................. .....131.58
Gordon.......................3.57................. .....217.66

Big question for me is, with doors open for some specialists already, what should be the cut-off for relif pitchers & the HoF? Mariano Rivera is a no-brainer; but do any of these guys. not admitted warrant a second look before opening floodgates for recent crops of relievers?

[brett]

Big question for me is, with doors open for some specialists already, what should be the cut-off for relif pitchers & the HoF? Mariano Rivera is a no-brainer; but do any of these guys. not admitted warrant a second look before opening floodgates for recent crops of relievers?

Wilhelm only started 52 games.
Gossage started 37 but it kills his career rates. He ahd a 146 ERA+ for his last 1200+ innings.

My view is that in general, a good reliever is not necessarily a good pitcher, but it is a guy who benefits from specialized situations and the ability to throw hard and not to be seen that much. Batters just don't get the chance to learn to deal with a guy who goes 70 innings. They are valuable because they are still better than a starter would be in his 4th time through the lineup.

So we should be very careful. How you get Fingers at a 2.78 adjusted ERA and Quiz at 2.81 is beyond me. Do you just use WHIP? Quiz had an ERA onlu 68.5% of his leagues (park adjusted) ERA. Fingers was 84% of his leagues adjusted ERA.

And even with you having the two about equal in adjusted ERA, how does Fingers end up with a 325 score and Quiz 95?

I would not put either one in the hall of fame, but I would rank Quiz higher.
I would not put Sutter in.

Rivera is so far ahead of all other relievers that we need to remember that. He deserves to be in without question. All the others raise questions, but I would add Wilhelm, and Eckersley (hybrid). That's my 3. I can see Gossage going in because excluding his time as a starter his stats are much more impressive. I think he had a 156 ERA+ through a stretch of over 1000 innings.

Rivera-so far ahead of the rest there is no comparison

Wilhelm
Eckersley

Gossage-right on the border

I might put Hoffman in but he only matched Quisenberry in ERA+ and through the same number of innings, and with the advantage of shorter appearances. Hoffman could get in more for a super longevity pass for me.

[leewileyfan]

Wilhelm only started 52 games.
Gossage started 37 but it kills his career rates. He ahd a 146 ERA+ for his last 1200+ innings.

Yes, I'm aware of Wilhelm's starts, as I alluded to in my post. I tried to make it clear that, in the case of Fingers & Wilhelm, I selected category that would present each in the best light.


My view is that in general, a good reliever is not necessarily a good pitcher, but it is a guy who benefits from specialized situations and the ability to throw hard and not to be seen that much. Batters just don't get the chance to learn to deal with a guy who goes 70 innings. They are valuable because they are still better than a starter would be in his 4th time through the lineup.

I'm guessing that over the last 20 years or so there have been young kids, all too soon groomed for roles prematurely; and, if they get caught up in constant adult oversight that tends to be fixated on trends, they may be stuck in a BB caste system.

It's a multiple self-fulfilling prophecy: You pamper starters with pitch counts; and you gat six inning pitchers. What do you do for that last one-third of a baseball game?

So we should be very careful. How you get Fingers at a 2.78 adjusted ERA and Quiz at 2.81 is beyond me. Do you just use WHIP? Quiz had an ERA onlu 68.5% of his leagues (park adjusted) ERA. Fingers was 84% of his leagues adjusted ERA.

How realistic is park adjustment for a submarine pitcher whose strong suits are control and keeping the ball low in the strike zone [or pot of it]?


And even with you having the two about equal in adjusted ERA, how does Fingers end up with a 325 score and Quiz 95?

Thanks for catching a typo. Man, I mistook the upsweep in "y" at the end of Quisenberry for part of the letter. It was a 1; therefore, Quisenberry = 195.91. The answer lies in career volume of work. Quisenberry, a finesse pitcher, was a virtual model for closers, with IP @ = or < appearances, while Fingers in relief pitched more innings per appearance and gave up relatively fewer hits in the process.

Fingers, in game context was dominant. Quisenberry was effective.


I would not put either one in the hall of fame, but I would rank Quiz higher.

IMHO, there are arguably a handful of no-brainers for the HoF in the relief corps; and three of them are Hybrids who also started:

Eckersly; Fingers; Wilhelm

#4 is Mariano Rivera; #5 is Billy Wagner; and #6 is Gossage.

The basic concern expressed in my post is that, with the gates now open, how many older guys with high value ratings warrant a second look before Percival, Hoffma, F. Rodriguez, and other current relievers [not even mentioned] get ushered in first.

Dick Hall? Sparky Lyle? Reardon? McDaniel? Miller? Not stumping for these guys, just asking.


I would not put Sutter in.

You don't have to. He's in.


Rivera is so far ahead of all other relievers that we need to remember that. He deserves to be in without question. All the others raise questions, but I would add Wilhelm, and Eckersley (hybrid). That's my 3. I can see Gossage going in because excluding his time as a starter his stats are much more impressive. I think he had a 156 ERA+ through a stretch of over 1000 innings.

See above. We agree.

[leewileyfan]

When it comes to starting pitchers, my own evaluations [same format as for relievers] argue that HoF membership might include a number of old-timers not yet inducted along with some recent retireees who may or may not get first ballot consideration. In this instance, as a time-saver, I'm just listing them in rank order by their value or greatness rankings. If anyone wants the adjusted ERAs, I'll post them later.

Starter...........................Value/Greatness

W. Johnson...........................1479.19
Clemens................................1072.66
Maddux.................................1062.25
Alexander................................976.09
Grove.....................................964.12
Mathewson.............................918.01
Seaver...................................908.42
Hubbell...................................861.39
R. Johnson..............................748.23
Martinez.................................746.20
C. Young................................740.24
Spahn....................................731.07
Walsh.....................................686.90
Feller......................................630.13
Ryan......................................594.39
Vance....................................577.38
Lyons.....................................556.10
Schilling..................................537.35
Ruffing....................................533.05
Plank.......................................531.77
Mussina...................................530.18
K. Brown.................................521.14
Derringer.................................505.88
Joss........................................503.95
E. "W" Ford..............................498.77
Glavine....................................496.53
M. Brown.................................488.15
E. "D" Leonard...........................487.12
G. Perry...................................476.95
Gibson.....................................474.70
Warneke..................................472.09
Carlton....................................469.49
Palmer.....................................461.42
Roberts....................................446.33
Gomez.....................................442.97
Jenkins....................................427.56
Blyleven...................................426.13
Walters....................................421.19
Wilhelm....................................419.97
Eckersley..................................413.54
Marichal....................................406.53
C. Mays....................................402.90
Wynn.......................................388.53
Waddell....................................386.57
"D" Dean...................................378.38
Drysdale...................................378.05
Bunning....................................367.93
Koufax......................................350.39
T. Bridges.................................350.39
Sutton......................................340.85


..... at this level, with several HoF pitchers not yet mentioned, I see a cusp, with some non-inductees above the line and some HoF below the line:

Pierce........................................320. 82
Brecheen....................................318.01
L. Jackson...................................311.96
M. Cooper...................................300.98
_____________________________________________

Newhouser..................................297.97
Rixey..........................................291 .98
"S" J. Wood.................................279.25
French........................................276. 21
Harder........................................291. 82
Chandler.....................................273.8 5
Moyer........................................279.0 9

Just as an added note, while I expected that Emil "Dutch" Leonard would hold up quite well, I was a bit surprised at how high his rating came in.

If one takes a moments to look at the bookends of his career, 2 of 3 as a "kid" when they weren't quite sure what to do with him and 4 at the end as an older veteran, his relief work was outstanding in two hellacious hitting climates.

I haven't done a detailed study; but I doubt many starters had career W-L records that so far outstripped his losing teams' records.

I don't believe any knuckleballer, with the exception of Hoyt Wilhelm, was ever measurably better than "Dutch" Leonard.
They pitched in notably different offensive settings.

[SABR Matt]

For someone who has zero interest in reading about the statistical products about which people are customarily asking in this thread...you sure are contaminating it quite a bit, lee.

[leewileyfan]

For someone who has zero interest in reading about the statistical products about which people are customarily asking in this thread...you sure are contaminating it quite a bit, lee.

Remember about a month or two back, I was disinclined to post anything here, because I had deferred, as a non-intrusive poster, t your personal fiefdom, as I called it.

It seems that you too see this entire thread as your proprietary station, owned and operated, without mortgage.

What you just wrote above translates into: "If you are not a devoted fan of PCA, you had better not post here, for SABT Matt's infallibility reigns supreme.

However, this is a public forum and your thread is part of that forum. You have not responded to me here with anything constructive, intelligently critical or willing to critique observations with corrective data.

You claim I am contaminating your thread. I'd suggest I've given it a newer perspective for people to consider. I shall continue to do so, whenever I see fit.

If you have anything intelligent to say, please do so as well. [Note: I did NOT start this - you did].

[SABR Matt]

What I said was not "if you are not a devoted PCA fan, don't post here"...what I did say was "If you are not here to actively discuss the THREAD SUBJECT!!!...do not post here"

This isn't a thread meant to be a dumping ground everyone's metrics. If everyone starts talking at random, this thread will become even harder to navigate than it already does.

It's not a fiefdom...it's following the same damned rules that apply to every other thread here.

If you want to discuss what you think is WRONG with my approach...(in an informed manner)...please feel free. You don't have to be a "fan"...whatever the heck that means. You do have to contribute to a discussion of PCA and relevant related metrics.

[Matthew C.]

Matt, I don't see this as a PCA question but I wondered where to find data on this question:

Do better pitchers tend "inspire" more run support-consistently.


I bet it does happen from time to time, but I also bet there is no way to scientifically find out how much/how often.

Also - Matt, still looking for a response to the HOF question if you have time - is there a DNRA marker level in which you look for regarding the HOF? Is it near 100? Thanks!

[SABR Matt]

I would just like to point out that while lee accuses me of being the instigator...I have carried on having a completely rational and open exchange of ideas, and unlike the bad scientists running the International Panel on Climate Change who have recently been exposed, I make my corrections and methods...and in fact as much data as I know how to provide...freely available. When someone spots a mistake, has a better idea, or leads me to spot a mistake on my own, I say so publicly.

STLCards2 is a great example of a person who came at me with his cletes up three years ago in a very vociferous debate about the merits of Tom Glavine's HOF candidacy. I was opposed at the time. I stated my case. He rationally, and politely (!) stated his. I took a lot of heat from less restrained members of this community. I probably would have dug in my heels and gotten exclusively defensive if it weren't for the fact that STLCards2 presented his case so well that it forced me to think about the best way to address his points...directly resulting in a new and IMHO significantly improved (over original pitching PCA) pitching metric and a great sense of respect for his insight. He still challenges me...we have argued for quite a while over Nolan Ryan, for example. But he's not entering every conversation with a sneer and I'm more than willing to hear him out.

If you are suggesting that he's some dyed-in-the-wool groupie for me, then you obviously don't know anything about our history. He's not...he can think for himself and we've had many exchanges where he disagreed with me. There's a reason his approach works with me and some do not.

If you're suggesting I don't respond to criticism, think again. In my six years here at baseball-fever.com, I have fought very hard for ideas I believed in...been persuaded I was wrong...and changed my thinking accordingly. Many times. I got into it with several posters here over whether one should measure production rate per OUT or per PA...they convinced me that counting production per OUT was not the best way to go since the margin for performance is calculated per OUT so rating production per OUT is going to result in a double-whammy of over-fixation on out rate.

I argued over the proper placement of the margin vs. the placement of the replacement level used in some other metrics. They convinced me to pursue the concept of marginal analysis based on line-up simulations in future works.

I argued over how best to rate pitching production...baseRuns or linear weights...they convinced me to use baseRuns over my initial objections.

I argued that defensive DIFFICULTY was inherently different than defensive VALUE and they convinced me that some of the value in a play made is in making a proper estimate of how many other players would make the same play. Future incarnations of PCA will include an adjustment to defensive values based on play-by-play era estimates of the difficulty of the average play made at each position.

Most recently, the debate centered on whether to use situational splits in evaluating players...particularly pitchers. I started out thinking that split stats occurred with samples too small to be certain of their significance. They convinced me to include situational splits for the defense independent events in my next pitching analysis.

I fought over any number of players' placement on my all-time lists...the baseball historians (NOT fellow sabermetricians) convinced me to move particularly some of the older players up in my rankings citing real world mitigating factors that complicate the analysis of said players. Examples include Joe Jackson (especially regarding his defensive ability), Jimmy McAleer, Buck Ewing, George Sisler and a handful of other players from the 1880s.

Approach me fairly, openly and in an informed manner...make your case without becoming confrontational and you will find me more than willing to listen and consider what you have to say. My history and my track record speak for themselves.

[SABR Matt]

I bet it does happen from time to time, but I also bet there is no way to scientifically find out how much/how often.

Also - Matt, still looking for a response to the HOF question if you have time - is there a DNRA marker level in which you look for regarding the HOF? Is it near 100? Thanks!

As the marker scores have been slightly revised due to the same-context analysis and the corrected exponent in all of my pythag calculations, I needed time to re-evaluate my leaderboard. However, having looked at all of the pitchers I would estimate that my HOF marker cut-off is around 100 +/- 10...(I have stopped thinking of the HOF as an in or out system even though that's how it works in the real world...I think of the hall as in, grey area, probably out and definitely out at this point)

I go back and forth about the worthiness of several of the pitchers in the 80s and 90s in marker score.

[Matthew C.]

He's not...he can think for himself and we've had many exchanges where he disagreed with me.
.

Yes, many! :)

[leewileyfan]

For someone who has zero interest in reading about the statistical products about which people are customarily asking in this thread...you sure are contaminating it quite a bit, lee.

This was not a bit of your post, snipped out of context. It was your entire post. This, after I had the gall & temerity to post in your thread evaluations of pitchers, relief and starter, in the context of thread exchanges on same being currently exchanged.

NOT A SINGLE WORD in that post recognized either the content of my post, the context in which it was explained or the work that went into it. Nor was there a single syllable that challenged my findings.

Your post was the post of a total snot, a brat without manners and ultra-defensive of his metric which you later demanded be at the core of all posts here. You ranted in defense of the challenges you faced and the graciousness with which you humbly made adjustments.

You attempt to deflect my direct observation to you as if I were attacking 2 other posters as your "groupies." I never suggested either was a groupie.

Man up. Address the topics at hand, not your own contaminated defenses.

[SABR Matt]

I will be reporting you the moderators. Name calling here is never acceptable...I should know...I've been banned for less offensive posts than that.

If my intent in the original post to which you objected was not clear, then that is my fault. But, I was upholding the rules laid out here at BBF that threads not get hijacked...this isn't a thread for your statistics...it's a thread for mine...mixing the two is not going to help keep the conversation readable. And given the complete lack of respect you've shown for the hard work I've done in my own research efforts, I don't think you're the person who should be lecturing me about how I've treated you.

[leewileyfan]

What I said was not "if you are not a devoted PCA fan, don't post here"...what I did say was "If you are not here to actively discuss the THREAD SUBJECT!!!...do not post here"

The topic under current discussion was pitcher evaluations.


This isn't a thread meant to be a dumping ground everyone's metrics. If everyone starts talking at random, this thread will become even harder to navigate than it already does.

I don't find it hard to navigate at all. Your personal seeding of personal landmines does get tedious, though.


It's not a fiefdom...it's following the same damned rules that apply to every other thread here.

Except for civility and maturity and thoughtful, constructive exchange of ideas.


If you want to discuss what you think is WRONG with my approach...(in an informed manner)...please feel free. You don't have to be a "fan"...whatever the heck that means. You do have to contribute to a discussion of PCA and relevant related metrics.

I don't know enough about your metric to criticize it. I am not the type of poster who lurks on sites to ridicule the work of others, either. I just ignore those, or move on.

PCA is your metric. Other relevant, related metrics include mine [whether you like it or not].

All metrics that explore the same topic are related. You alone are in no position to determine their respective relevance. [Especially when you do not take the time to understand them]. You have never honstly challenged a single piece of data I've posted. I don't know why you seem threatened by my presence here.

I am certainly not threatened by yours.

[SABR Matt]

LOL...suggesting I'm professionally threatened by your metrics is pretty hilarious. Sorry, but no. I am annoyed, however. I've spent 8 years learning to analyze the game and taking on the best ideas of everyone with whom I have conversed and you're taking the field back twenty years with your little toys and then claiming authority over anyone else's work even though you've demonstrated absolutely no willingness to read the existing literature on sabermetric analysis and converse intelligently about said analysis (by intelligently, I mean from an informed position).

And I am annoyed that you complained about my not staying on topic in a thread about catchers and then you pull the same crap in a thread specifically called the PCA REQUEST LINE and then have the unmitigated gall to act as though you haven't gone off subject.

The mods can feel free to ban me...because this site is not fun so long as you continue to masquerade as an analyst...the gloves can come off, since I'm done with this crap. Your work is a JOKE...the fact that you cannot see why you're 20 years behind the leading edge in analysis is a sad statement about your own IGNORANCE.

Good day all. I would ask while you're banning me that you please close this thread so we don't have to see leewiley spouting his garbage and wrecking what had been a fun thread in years past.

[leewileyfan]

I will be reporting you the moderators. Name calling here is never acceptable...I should know...I've been banned for less offensive posts than that.

Read! I never called you a name. I referred to your content as that of a snot [sassy, disreptectful kid] and brat [perhaps a redundancy, not an epithet]. How adult of you!


If my intent in the original post to which you objected was not clear, then that is my fault. But, I was upholding the rules laid out here at BBF that threads not get hijacked...this isn't a thread for your statistics...it's a thread for mine...mixing the two is not going to help keep the conversation readable.

Funny. I've seen lists from STL on pitcher measures [metric] and brett [on various formulae and inputs]. I posted lists, with explanations for same, in the context of a discussion on pitcher evaluation. Maybe I should have prefaced it with my customarily humble FWIW.


And given the complete lack of respect you've shown for the hard work I've done in my own research efforts, I don't think you're the person who should be lecturing me about how I've treated you.

I have never dissed PCA. I have stated here that I have attempted to find it.

I have done Yahoo, Google and blog site searches; and after 8 pages of other headings for "PCA" I gave up.

After you have attacked me I lost all interest in PCA.

What about my hard work? You are extremely arrogant. [Report that].

[leewileyfan]

Approach me fairly, openly and in an informed manner...make your case without becoming confrontational and you will find me more than willing to listen and consider what you have to say. My history and my track record speak for themselves.

Been there. Done that. This very day, within hours. Look what it got me.

[AstrosFan]

I have done Yahoo, Google and blog site searches; and after 8 pages of other headings for "PCA" I gave up.

Don't fiddle with search engines. All you need to do is join the "baseball databank" Yahoo group. Once you're in, just go to the files section, and scroll until you see the one labeled "PCAManifest.pdf". You can read the entire work, as it was written then though. Matt updates all the time.

[brett]

Thanks for catching a typo. Man, I mistook the upsweep in "y" at the end of Quisenberry for part of the letter. It was a 1; therefore, Quisenberry = 195.91. The answer lies in career volume of work. Quisenberry, a finesse pitcher, was a virtual model for closers, with IP @ = or < appearances, while Fingers in relief pitched more innings per appearance and gave up relatively fewer hits in the process.


Quiz had 1043+ IP in 674 appearances-all as a reliever (1.55 per relief appearance).
Fingers had 1505+ IP in 907 games as a reliever (1.66 per relief appearance).

Fingers pitched in the most pitcher friendly park for most of his career.
Quiz had a 1.175 WHIP to 1.156 for Fingers in that low environment.

[Matthew C.]

Rivera-so far ahead of the rest there is no comparison

Wilhelm
Eckersley

Gossage-right on the border

.

I agree 100% with this assessment.

[leewileyfan]

Quiz had 1043+ IP in 674 appearances-all as a reliever (1.55 per relief appearance).
Fingers had 1505+ IP in 907 games as a reliever (1.66 per relief appearance).

Volume of relief appearances helps widen the gap in BF and challenges faced per appearance. Fingers made half-again more relief appearances than Quisenberry; and he pitch @ 7% more per appearance.

Fingers pitched in the most pitcher friendly park for most of his career.
Quiz had a 1.175 WHIP to 1.156 for Fingers in that low environment.

Let's look at some facts:

1. Fingers did break in in 1969 the season just following the MLB hitting plunge scraping bottom. However, when he started specializing as a reliever, most of his seasons averaged a hitting climate of @ .258/.325/.375 with @ 697 OPS.

2. Quisenberry broke in in the season of his most hostile batting atmosphere, piching 40 innings in a .270/.334/.408 climate @ .743 OPS. In his final season Fingers pitched 55+ innings in a .261/.327/.406 atmosphere @ .733 OPS. The offensive climate edge goes slightly to Quisenberry; but the consistency and durable effective longevity of Fingers wins overall.

Batting during Fingers' career rose, especially in the A.L. with the DH. Fingers pitched in OPS seasons > .700 in 8 of his seasons. Quisenberry faced such a climate 9 times.

Look at both men's careers for season-by-season WHIP; and I don't believe any reasonable reader would debate the consistent superiority and work rates dominance of Fingers. Almost half of Quisenberry's career was post-peak decline.

3. Opposing batters fared as follows against the two:

Against Quisenberry : .267/.294/.393 @ .657 OPS

Against Fingers: .235/.292/.340 @ .632 OPS

4. Quisenberry wins big, as a submariner, inducing GDP at a better rate than Fingers. Fingers wins on powe numbers with K/9 @ 7.8 & BB @ 2.5 vs Quisenberry's K's/9 @ 3.3 & BB @ 1.4 [excellent control].

[Ubiquitous]

Gentlemen,
Please stick to discussing and debating the merits of the baseball topic and refrain from discussing the personalities of the individual posters here on baseball-fever. Name calling is never acceptable and rarely accomplishes much. Please carry on discussing baseball, any questions feel free to PM me but please do not try to start a conversation about this in this thread. Talk baseball in this thread.

thank you.

[leewileyfan]

Don't fiddle with search engines. All you need to do is join the "baseball databank" Yahoo group. Once you're in, just go to the files section, and scroll until you see the one labeled "PCAManifest.pdf". You can read the entire work, as it was written then though. Matt updates all the time.

Astros Fan:

Thank you. I followed the procedure and successfully downloaded the PCA Manifest [pdf] file, which I have read.

[AstrosFan]

Astros Fan:

Thank you. I followed the procedure and successfully downloaded the PCA Manifest [pdf] file, which I have read.

Good deal.

[SABR Matt]

I gave the same instructions 3 or 4 different times to you, lee. But whatever.

[leewileyfan]

I gave the same instructions 3 or 4 different times to you, lee. But whatever.

A good silence is better than a bad speech.

Astros Fan was the single poster who made it very clear that I had to join a specific group once into the general Yahoo fold, as if I were a brand new sign up.

In my previous attempts, as a Yahoo member participating in several groups and functions, I easily accessed whatever I wanted with no such added sign up.

Thanks again, Astros Fan.

[brett]

Let's look at some facts:

1. Fingers did break in in 1969 the season just following the MLB hitting plunge scraping bottom. However, when he started specializing as a reliever, most of his seasons averaged a hitting climate of @ .258/.325/.375 with @ 697 OPS.

2. Quisenberry broke in in the season of his most hostile batting atmosphere, piching 40 innings in a .270/.334/.408 climate @ .743 OPS. In his final season Fingers pitched 55+ innings in a .261/.327/.406 atmosphere @ .733 OPS. The offensive climate edge goes slightly to Quisenberry; but the consistency and durable effective longevity of Fingers wins overall.

Batting during Fingers' career rose, especially in the A.L. with the DH. Fingers pitched in OPS seasons > .700 in 8 of his seasons. Quisenberry faced such a climate 9 times.

Look at both men's careers for season-by-season WHIP; and I don't believe any reasonable reader would debate the consistent superiority and work rates dominance of Fingers. Almost half of Quisenberry's career was post-peak decline.

3. Opposing batters fared as follows against the two:

Against Quisenberry : .267/.294/.393 @ .657 OPS

Against Fingers: .235/.292/.340 @ .632 OPS

4. Quisenberry wins big, as a submariner, inducing GDP at a better rate than Fingers. Fingers wins on powe numbers with K/9 @ 7.8 & BB @ 2.5 vs Quisenberry's K's/9 @ 3.3 & BB @ 1.4 [excellent control].

Just for reference, for every approximately 10 points increase in league OPS you get about a 0.1 run per game increase (a little more actually).

The difference between Quiz' .733 and Fingers' .697 is a difference of about 0.4 runs per game, or a 10% increase in run environment.

Basically that means that if you take Fingers' ERA and multiply it by 1.1 you would convert it to Quiz' setting, or take Quiz' ERA and divide it by 1.1.

And I can see putting Fingers higher, but I would have to keep both out of my hall.

[brett]

Matt, so Quisenberry has a DNRA+ of about 122 versus a real ERA+ of 146

Gossage and Fingers show almost no drop from real to DNRA+ and Sutter has a much smaller drop.

What are the main factors than allowed Quiz to put up a 146 pitching like a 122? KC tended to have a lot of fast outfielder, and Frank White at second, but overall I don't think they were a great fielding team from '80-'85.

It could be a) good defenses; b) lucky/good splits; other?

Its not splits.
split relative OPS+
bases empty=88
men on 113

His unearned run rate is not bad, a little under 10%. I think that's low for a reliever.

The only thing I can come up with is that his relative OB% allowed was much better than his relative slugging allowed and that maybe this doesn't get picked up by baseruns.

His OPS+ allowed is about 85, but his relative OB% allowed is 0.907 while his relative slugging is 0.945 (about half as much "better" than average") Still, its clear that his ERA+ is much better than his OPS+ allowed, even using a higher OB% weighting.

[SABR Matt]

ERA is not a good stat to use for relief pitchers. Closers in particular face the potentially biasing factor of runs that never get a chance to score. If you blow a lead in the bottom of the 9th inning, unless the final hit is a home run, you can't score more runs than the amount needed to win the game. When you combine that with closers and set-up men allowing the preceding reliever's inherited runners to score and not getting blamed for that damage, you end up with a two-sided problem that renders ERA a bad tool for the reliever.

Why is Quisenberry's DNRA+ closer to his inverted OPS+ than his ERA+? Most likely because Quisenberry was asked to take over in a lot of games where the opposing team already had runners on base and he allowed some of those guys to score (showing up in OPS+ and not ERA+) and because on the rare occasions when he got beat up, some of them occurred in the bottom of an inning and where the run expectancy might have been worth 2 or 3 or 4 runs, he only actually got charged with 1 or 2.

The problem we have here is that Quiz (and all relievers) just doesn't have enough total innings where ERA (with its' many potential biases for short relievers) is likely to converge to the true level of a pitcher's performance.

[brett]

One thing I considered is that Gossage, Sutter and Fingers were much better K pitchers so they could save an unearned run from a runner who was on third when they entered the game with fewer than 2 outs by avoiding contact.

The problem with that is that Quiz had the lowest ratio of unearned runs allowed to total runs allowed among those 4 pitchers. That throws out the inherited runner not charged theory.

He also had the lowest blown save percentage of the 4 I am pretty sure based on memory, so the maxing out theory also doesn't seem to hold.

The other possibility with DNRA+ might be that his ballpark was particularly suited to his style and so his ERA+ gets a break while DNRA+ picks that up.
Quiz was not any better at home than on the road though.

Actually with Fingers and Gossage, their DNRA+ might not drop off as much in part because they started a chunk of games in their careers.

What is peculiar about Quiz was that his BABIP allowed by season corresponds directly with his career path.
BABIP and ERA+
'79: .274 136
'80: .276 131
'81: .274 207
'82: .254 159
'83: .241 210
'84: .252 153
'85: .298 175
'86: .308 153
'87: .301 166
'88: .353 72
'89: .291 137
'90: .387 27

'82-'84 were clearly his prime years. In '80 he was still developing consistency and by the end of '85 there was talk that he had lost his touch.

And we can blame sample size, but he pitched 400 innings from '82 through '84.

It's weird but to me it looks like a guy who's success and failure were due to BABIP allowed.

[SABR Matt]

What the heck are you talking about? If a pitcher gives up someone else's inherited runners, it does *NOT* go into his column as an unearned run. I have no idea where you came up with that. It goes into the previous pitcher's run column (and if it was earned...his earned run column)...no mention of the run is ever placed in the replacing pitcher's stat sheet.

[brett]

What the heck are you talking about? If a pitcher gives up someone else's inherited runners, it does *NOT* go into his column as an unearned run. I have no idea where you came up with that. It goes into the previous pitcher's run column (and if it was earned...his earned run column)...no mention of the run is ever placed in the replacing pitcher's stat sheet.

Oh yea. I never thought that far. Obviously a pitcher who leaves the game and is charged with an earned run is also charged with a run. So where do we get stats on inherited runners?

[SABR Matt]

http://www.baseball-reference.com/players/q/quiseda01-pitch.shtml

Scroll down to "reliever pitching".

WOW! I believe, ladies and gentlemen, that we have found the smoking gun.

In his career, Quisenberry allowed *38%* of his inherited runners to score. Given that the average strand rate in major league baseball is 71% today...in an era of higher run-scoring rates...and likely higher than that back in the early 80s...I believe it is safe to say that because of Quiz's inability to get the K, he allowed way too many inherited runners to score.

[Matthew C.]

http://www.baseball-reference.com/players/q/quiseda01-pitch.shtml

Scroll down to "reliever pitching".

WOW! I believe, ladies and gentlemen, that we have found the smoking gun.

In his career, Quisenberry allowed *38%* of his inherited runners to score. Given that the average strand rate in major league baseball is 71% today...in an era of higher run-scoring rates...and likely higher than that back in the early 80s...I believe it is safe to say that because of Quiz's inability to get the K, he allowed way too many inherited runners to score.

Since this is the case, I would say any chance of Quiz being a personal "HOFer" are toast. An "All-time" reliever allowing almost 25% more inherited runners to score than average? I wonder how this compares to the other 2nd tier relievers (Smith, Hoffman, Franco, Sutter, Fingers, Wagner, Gossage)

[SABR Matt]

Smith: 28%
Gossage: 33%
Hoffman: 21%
Franco: 32%
Sutter: 31%
Fingers: 29%
Wagner: 28%

The longer you have to pitch, the harder it is to keep your IRS% low...that's observation number 1...observation #2 is that high-K relievers prevent IRS better than low-K relievers. Observation #3 is that Quisenberry is rather severely worse than any of these guys.

[SABR Matt]

Just for fun...if we correct Quisenberry's ERA by charging him with the difference in runs allowed between his IRS and the expected IRS of 30%...

IR = 523
IRS = 201
XIRS = 157
DELTA = 44 (!)
New ER Allowed = 364
New ERA = 3.14
New ERA+ = 128

Feeling better about the DNRA score yet?

[brett]

[url]I believe it is safe to say that because of Quiz's inability to get the K, he allowed way too many inherited runners to score.

I like it when the world makes sense. I always thought that his low K rate was hurting him in those key situations.

[brett]

Just for fun...if we correct Quisenberry's ERA by charging him with the difference in runs allowed between his IRS and the expected IRS of 30%...

IR = 523
IRS = 201
XIRS = 157
DELTA = 44 (!)
New ER Allowed = 364
New ERA = 3.14
New ERA+ = 128

Feeling better about the DNRA score yet?

I didn't really have a problem with DNRA. I knew that Quiz had similar WHIPS and run environments to Fingers, Gossage and Sutter so I knew that he had to have a similar level of effectiveness, (plus leewileyfan had Quiz with about the same adjusted ERA as Fingers) but he has that 146 ERA+ that stood out from the others. Its still an anomalous ERA+ though.

[brett]

Smith: 28%
Gossage: 33%
Hoffman: 21%
Franco: 32%
Sutter: 31%
Fingers: 29%
Wagner: 28%

The longer you have to pitch, the harder it is to keep your IRS% low...that's observation number 1...observation #2 is that high-K relievers prevent IRS better than low-K relievers. Observation #3 is that Quisenberry is rather severely worse than any of these guys.

Care to take a stab at Hoffman's DNRA+ at this point? I thought he was overrated, especially since he has a lot of support to get in, but he might have a case. He just went over 1000 IP last year I think.

[SABR Matt]

Trevor Hoffman's DNRA+ as of the end of the 2005 season was 144. It may have edged down very slightly since then, but I doubt it was by much.

It should be noted that Hoffman allowed only 70 IRS when he should have allowed 102, so that's 32 earned runs he saved his teams...His ERA+ is already 147...accounting for the runs he saved, it goes up to about 170...the counterbalance is that he probably gave up some baserunners who never got a chance to score due to the bottom of the last at bat problem I mentioned earlier.

[brett]

Trevor Hoffman's DNRA+ as of the end of the 2005 season was 144. It may have edged down very slightly since then, but I doubt it was by much.

It should be noted that Hoffman allowed only 70 IRS when he should have allowed 102, so that's 32 earned runs he saved his teams...His ERA+ is already 147...accounting for the runs he saved, it goes up to about 170...the counterbalance is that he probably gave up some baserunners who never got a chance to score due to the bottom of the last at bat problem I mentioned earlier.

That doesn't sound like a bad hall of fame case. I'd still have him below Gossage who I have as a borderline but not illegitimate hall of famer.

I think that when compared season by season, Gossage probably would look better than Hoffman in PCA.

I've got Rivera as a HOFer, and Eck and Wilhelm, while Gossage I probably wouldn't have put in, but still, season by season he looks a lot better because his career ERA+ is trashed by his early starter years. He was like 127 overall, but 146 for 18 years and 1200+ innings to end his career, and 152 for over 1100 from '77 to '92.

[brett]

When possible Matt, could I see a comparison of Gehrig, Hornsby, Collins, Schmidt and Henderson using PCA wins, marker, offensive and defensive.

This is a group of 5 I have very close. If you have the time I'd appreciate seeing them season-by-season.

[Second Base Coach]

I would bill him for that kind of work...

[Jsquared83]

My apologies if this thread is dead but how does Andy Pettitte and Mark Teixeira's career PCA numbers look through '09, if that information is fully available? Thanks in advance.

[White Knight]

SABR Matt, whatever happened to you? 8,733 posts, yet none since March?

[RuthMayBond]

SABR Matt, whatever happened to you? 8,733 posts, yet none since March?
ElHalo hasn't posted in over a YEAR AND A HALF, and there are still only NINE who have more posts than him :bowdown:

[SABR Matt]

Reports of my demise are slightly exaggerated...I am, however, in the later stages of my master's degree in meteorology, and thus not rolling i free time...LOL

[Honus Wagner Rules]

Matt! Nice to see you are back.

[nerfan]

ElHalo hasn't posted in over a YEAR AND A HALF, and there are still only NINE who have more posts than him :bowdown:

Including you. I've come to believe that RMB actually stands for Really Mad BBFPoster.

[RuthMayBond]

Including you. I've come to believe that RMB actually stands for Really Mad BBFPoster.Everyone else has known that for years :crazy

[Matthew C.]

hopefuly some time will clear up for you- this place is not as exciting without you!

[SABR Matt]

I occasionally cause trouble, I know...a fact I have been working on of late...but I do appreciate the support. :)

You are one of the people I bring up when talking about how productive and fun a well-reasoned debate can be for me...how much I feel like I can learn when I encounter someone willing to put with my Socratic debating style...LOL

[Matthew C.]

I occasionally cause trouble, I know...a fact I have been working on of late...but I do appreciate the support. :)

You are one of the people I bring up when talking about how productive and fun a well-reasoned debate can be for me...how much I feel like I can learn when I encounter someone willing to put with my Socratic debating style...LOL

I never mind a debate...until it becomes personal. And never once have any of your posts been a personal attack to me. Can't say the same for dozens of others here.

[SABR Matt]

Yeah...I've made the mistake of getting personal with a few posters here - generally in response to dismissive tone or personal attacks directed at me or someone else here...but this kind of exchange has led me to the conclusion that the best response to such things online is dead silence.

And I neved felt any great need to personally attack you in my own (or a friend of mine's) defense because you have never once been dismissive of me, nor needlessly attacked anyone here. As I've said...talking to you about this great game has generally led me to an improved understanding of the game and of my own views...that's a rare and valuable thing.

[SABR Matt]

Incidentally, STL...I was just in a discussion revolving around the question of whether Jim Palmer was an ace in the historical context...or whether the historically brilliant Oriole defense of the 70s elevated a very good, but not great pitcher to an undeserved status as cultural icon. I am curious what you believe?

[Matthew C.]

Incidentally, STL...I was just in a discussion revolving around the question of whether Jim Palmer was an ace in the historical context...or whether the historically brilliant Oriole defense of the 70s elevated a very good, but not great pitcher to an undeserved status as cultural icon. I am curious what you believe?

From everything I have seen, he was a mid 30's all-time guy on his own merit that looks like an early-mid 20's guy without putting his defense into context. It looks like about half of his BABIP prevention was his own doing and about half his defenses. I am still not sure what to do when a guy like Palmer pitches to contact on purpose because he knows his defenses are good. Seems like he deserves a little credit for that too -especially if it limited HRs and BBs.

By the way, how is the DNRA+ stuff going? Any publications in the work?

[SABR Matt]

No time to pursue publications in baseball research unfortunately...

DNRA+ hasn't changed much since the last time we spoke either, although I have played around with a few potential tweaks to the system, mostly dealing with strength of schedule and situational statistics.

When you say mid-30s pitcher...do you mean mid-30s all time? Because based on his DNRA profile, I've got Palmer around the 90th best pitcher of all time...which would not qualify him for my hall of fame cut-off...but I am curious where you place him on a similar list.

[Matthew C.]

No time to pursue publications in baseball research unfortunately...

DNRA+ hasn't changed much since the last time we spoke either, although I have played around with a few potential tweaks to the system, mostly dealing with strength of schedule and situational statistics.

When you say mid-30s pitcher...do you mean mid-30s all time? Because based on his DNRA profile, I've got Palmer around the 90th best pitcher of all time...which would not qualify him for my hall of fame cut-off...but I am curious where you place him on a similar list.

TZ and the like see Palmer's defenses saving about 150 or so runs for him. Despite TZ limitations, that seems about right - which would put him from about 80 WAR to 65 WAR, if you are into WAR at all (I know you are not). That all seems reasonable to me. For a guy to be at 4,000 Ip with a 127 ERA+ on the surface to be a 90th best pitcher on his own merits... it just seems too big of a gap for any defense to be responsible for on its own. Are there other non-defensive factors that DNRA+ is picking up on?

Also, on your last DNRA marker list, you had Palmer around #25 or so since the early 50's. How does a guy who you have as #25 since the 50's end up near #90 all-time? Surely there aren't 65 pre-integration pitchers better than Palmer.

And you mention strength of schedule:

Here is something I dug up when somebody asked me about SOS and modern pitchers...

% of teams faced over .500:

Clemens-46%
Maddux-51%
Johnson-45%
Mussina - 48%
Martinez-50%
Glavine-53%
Schilling-51%
Smoltz-51%
Brown-47%

So the gap from worst to best is 8%. Is this negligible or impactful? Should the difference between Maddux and Clemens be ignored or can it play an impact on a "who is better" debate? If the 6% that separates them is negligible, how much of a gap is needed to matter?

[SABR Matt]

90th is a typo...that should have been 60th (what do you get when you cross a visually impaired person with a laptop instead of his usual desktop workstation?)

60th qualifies you for the Hall if you go with the "large hall" interpretation. His high IP count gets him enuogh total value to get fairly high on the career total WAR...but you run into a problem when you convert that to HOF Marker as I've traditionally done (wins + wins above average for each season summed up) because Palmer has only 4 seasons that would qualify as particularly dominant relative to average. Not...exactly...a big HOF resume there...

As I do not currently have access to my DNRA database (on vacation) I don't recall whether Palmer was 25th on the DNRA+ rankings, the DNRA wins created list...or the DNRA Marker list. Do you have that info?

At any rate, regarding strength of schedule, in a universe where the difference between being a really bad team...and being a world champion...is 20%, I would say that if the difference between the most difficult and least difficult schedules for various pitchers was 8%...that would be significant to about the same degree as team defense (read: very significant in extreme cases, and significant enough not to ignore in most cases...especially year to year).

If you recall, a couple of years back, I actually used the game log database to calculate the strength fo offenses faced by each starting pitcher (after all, we don't care about the total team strength of the opponent...just the bats) and found that even on the CAREER level, there were pitchers in my top 100 list all time by PCA that have OW% faced as high as .525 and as low as .471...I'd call that significant. That is one of the major things I think is missing from current pitching projections...along with umpire bias, weather affects, and park effects limited only to the thing the park actually impacts for a pitcher that is not already accounted for by removing the team defense from the equation - home runs and possibly K/BB ratio (if there's a visibility issue like at early Safeco Field)...all of those things matter in evaluating a pitcher. More than we realize. Especially when looking at individual years of data.

[Matthew C.]

Here is what you posted on July 19th, 2009. I believe that pitcher offense and all 1999 data is excluded if my memory serves me correctly:

Roger Clemens 305.9
Greg Maddux 236.5
Randy Johnson 225.2
Pedro Martinez 199.0
Nolan Ryan 198.4
Tom Seaver 167.9
Gaylord Perry 165.1
Kevin Brown 154.0
Bert Blyleven 153.1
Mike Mussina 144.2
Steve Carlton 143.5
Curt Schilling 129.8
Bob Gibson 122.5
Fergie Jenkins 121.6
Juan Marichal 118.4
Don Sutton 117.6
John Smoltz 115.7
Tommy John 114.8
Tom Glavine 113.1
Phil Niekro 112.1
Dennis Eckersley 111.2
Mariano Rivera 111.0
Sandy Koufax 104.4
Dave Stieb 103.5
Jim Palmer 98.4
Don Drysdale 96.7
Bret Saberhagen 95.9
Rick Reuschel 95.3
Jim Bunning 94.5
David Cone 94.5

I agree 100% that strength of offense faced is a huge hole that needs to be plugged.

[SABR Matt]

I don't include offense for pitchers because I don't know the best way to handle the notion that AL pitchers never get to hit. No offensive contribution is not the same as average offensive contribution...or even replacement level. It could wind up biasing the data rather badly.

I believe these are Marker wins based on the range...so I've got Palmer 25th in marker wins...not an unrealistic listing, I don't think. Though I do wonder how their strengths of schedule varied and how that might impact the final ratings.

[Matthew C.]

I don't include offense for pitchers because I don't know the best way to handle the notion that AL pitchers never get to hit. No offensive contribution is not the same as average offensive contribution...or even replacement level. It could wind up biasing the data rather badly.

I believe these are Marker wins based on the range...so I've got Palmer 25th in marker wins...not an unrealistic listing, I don't think. Though I do wonder how their strengths of schedule varied and how that might impact the final ratings.

I agree that finding out "how" to deal with pitcher offense is very difficult. But I am not too worried about the AL issue. When looking at the entirety of the league's history, a huge majority of pitchers who would be affected played prior to the DH era, or played most of their time or enough time in the non-DH NL. When we narrow it down to pitchers we would care about ranking (the top 200 or so), how many were DH era, AL only/mostly pitchers? Not too many compared to the rest. We may not know how Stieb, Mussina, Guidry, or Saberhagen are affected by including pitcher offense, but we do know the impact is statistically significant for Gibson, Drysdale, W. Johnson, Ruffing, Ferrell, Caruthers, Lemon, Wynn, Glavine and many others on one side and Sutton, Koosman, Koufax, Coveleski, Willis, Faber, Galvin and many others on the other side.

It does make you wonder how some proven (from their NL years) poor hitters' (such as R. Johnson, N. Ryan, and C. Schilling) WAR,s DNRA+s, etc. would be affected if they didn't get to dodge pitchers for large chunks of their careers.

Clemens is the only HOF level guy I can think of who was a good hitter during an NL stint that would have benefited from no DH.

Bottom line: I'd rather screw over a Jack Morris or two (or give them a tiny boost) when we don't know, than to screw over or artificially enhance a whole lot of other guys when we know it impacted the game.

[SABR Matt]

Eeehhh...I can understand that philosophy, but I don't share it. My standing policy regarding sabermetrics is the scientific one...if a comparison literally cannot be made because the subjects operate under different realities...I don't make the comparisn. The vast majority of pitchers are far-sub-RL hitters...there are differences that matter...but because perhaps 20% of the pitching population is infiltrated with missing data (even if they did hit some of the time...a good chunk of pitchers are missing years to the DH-era AL)...trying to make some sort of definitive conclusion about rankings under those circumstances that includes the offense is potentially misleading enough for me not to want to touch it.

[Matthew C.]

Fair enough - a hitting bonus/deduction can always be made subjectively seperate from sabermetric rankings when considering HOF, etc. Same with postseason. I am sure you would agree that finding an objective way to measure postseason success is impossible, but should be considered subjectively when considering HOF criteria.

[SABR Matt]

Post-season analysis is far from impossible. We know what the strength of schedule is in the post-season based on how good the teams were during the same regular season. We know the play by play for every post-season game that's ever been played. Why exactly can't we include post-season stats into the big picture? The only argument you might make that excludes the post-season is that it's not a pitcher's fault if his team sucked all the time and he never got to the post-season...similar to the batting problem. This is why I don't personally add post-season stats to the analysis...but I would not describe it as impossible to do so if you choose.

My approach to the post-season, to seasons lost to war or team incompetence (see: Grove getting delayed for promotion for three years for...no real reason...same for Edgar Martinez etc), to pitcher batting, and to intangibles like reputation for leadership qualities, clutch performance and such is to do much like what James did and apply a subjective tweak to the ratings when I feel there is real reason to do so.

[Honus Wagner Rules]

Matt how is Albert Pujols' defense according to PCA over the past few seasons?

[SABR Matt]

Well I can tell you that PCA finds absolutely no evidence to support the claim made by UZR that Pujols was a +25 (!!!!) run defensive first baseman. Above average? Yes. Better than any defensive first baseman since...ever? Not on your life, Paco. That number appears out of nowhere...is surrounded by much more reasonable seasons...and then disappears.

I have Pujols running short-hand defensive PCA wins like:

2006: 3.2 (probably the PCA GG winner in 2006)
2007: 3.0
2008: 2.8 (still above average by quite a bit)
2009: 1.6 (evidence exists that his recurrign hip issues may have begun ruining his range)
2010: 1.3 (below average now and decaying rapidly)

This is short-hand only...so the full analysis might be more optimistic...but I doubt it will look like his Fangraphs card...wow. This is why I hate UZR with the fire of a thousand stars when viewed in single-season snippets. PCA has generally proven in my side by side comparisons to be much more stable.

[Matthew C.]

The only argument you might make that excludes the post-season is that it's not a pitcher's fault if his team sucked all the time and he never got to the post-season...similar to the batting problem. .

This is exactly why. And the issue of more and more playoff rounds as the eras have passed.

[Matthew C.]

Well I can tell you that PCA finds absolutely no evidence to support the claim made by UZR that Pujols was a +25 (!!!!) run defensive first baseman. Above average? Yes. Better than any defensive first baseman since...ever? Not on your life, Paco. That number appears out of nowhere...is surrounded by much more reasonable seasons...and then disappears.

I have Pujols running short-hand defensive PCA wins like:

2006: 3.2 (probably the PCA GG winner in 2006)
2007: 3.0
2008: 2.8 (still above average by quite a bit)
2009: 1.6 (evidence exists that his recurrign hip issues may have begun ruining his range)
2010: 1.3 (below average now and decaying rapidly)

This is short-hand only...so the full analysis might be more optimistic...but I doubt it will look like his Fangraphs card...wow. This is why I hate UZR with the fire of a thousand stars when viewed in single-season snippets. PCA has generally proven in my side by side comparisons to be much more stable.

I agree that PCA seems more stable from year to year, but I will state for the record that UZR also saw a significant slowdown in 2009 and 2010 for Pujols.

[SABR Matt]

Yes it does...it just doesn't make any logical sense in 2007...how the heck can you even save 25 runs above average at first base? That just seems very...very improbable. The best defensive first basemen of all time by PCA save 20-22 runs in their very best years above average (they get 4-4.2 win seasons when the average 162 game win total at first is 1.7...but PCA has always linked a win with 8.5 runs or so...not 10)...and I doubt that Pujols was ever one of the greatest defensive first basemen of all time, even for a single season. Good, yes...historically great? Nada.

[SABR Matt]

Agreed that the inequity in psot-season playing time causes problems...you could theoretically try something like WPA, but with each post-season game counting for more the fewer post-season games there are in a given season...but it's probably not worth the effort. Better to stick with presenting the statistics for the post-season separately and using them to subjectively adjust a player's value if there is a strong case for or against that player in the data. I do not, for example, make any tweaks to the rankings unless the subjective evidence is very strong...preferring to keep things as objective as possible.

[SABR Matt]

Incidentally, if anyone is interested in why I think PCA might be better able to stabilize the defensive numbers...I have three theories.

1) Unlike UZR, PCA treats defenses as a set of units that do not really interact with each other much, but that are internally collective in their efforts. UZR assumes that each fielder is unto an island...that because zone overlaps are rare (how many plays can you count where more than one player can conceivably make the play?), the inteconnectivity of the positions is negligible. I consider this position to be absolutely false. I believe that Albert Pujols changes how he plays first depending on a number of factors relating to his other infielders. I would note, for example, that in 2007, Pujols' mates at 2B were about 5 runs below average, whereas in 2006, they were about average exactly. Players change how they play in many situations based on the ranges of their teammates.

2) PCA is normalized so that the standard deviation of production is constant throughout time. This is, IMHO, important because it removes the assumption that our definition of average plays made by zone (defined differently each year) should really change with time and in so doing, compares players against one standard based on a much larger sample size of all performances at their respective position. UZR has, as a potential flaw, IMHO, the problem of small sample. You may not think so, given the huge number of plays that get made at each position, but keep in mind...we're not actually comparing play to play...we're comparing player to player within a single league. In the NL, there were only 20 or so first basemen with enough innings to qualify as significant to the definition of what is average performance at the position (and determine the spread of that performance). In essence, PCA assumes our ignorance...we only played the season once. If we played the season a million times with every player contributing the same skill level in any siulation...what would the average win value of a person's defense be?

3) PCA has the added potential benefit of accounting for the plays which actually do involve multiple people in a more logically consistent way. UZR makes an attempt to quantify (on an individual basis) things like arm elements and DP turning ability...but PCA, since everything is cooperative within a defensive unit, can attribute credit for group-effort plays intuitiviely.

All of this leads to a system that, I believe, is more likely to possess predictive skill than UZR.

[Matthew C.]

Agreed that the inequity in psot-season playing time causes problems...you could theoretically try something like WPA, but with each post-season game counting for more the fewer post-season games there are in a given season...but it's probably not worth the effort. Better to stick with presenting the statistics for the post-season separately and using them to subjectively adjust a player's value if there is a strong case for or against that player in the data. I do not, for example, make any tweaks to the rankings unless the subjective evidence is very strong...preferring to keep things as objective as possible.

+/- and TZ agree with UZR with Pujols being an all-time defensive great at 1st. Is there something they all share in common that is causing the issue in your eyes?

[SABR Matt]

Well...they're all based on the same basic idea of counting the balls in a player's static zone that he missed and caught and comparing that to a league average for his position...a system I have never been a big fan of...but it could well be that PCA is legitimately missing a tremendous season by Pujols...I jsut never really saw him as that kind of fielder...he was always a good one...but not the first one I thought of in his own league most years. I have a hard time believing stats that show a player having one gigantic career year on defense surrounded by much more modest stuff. *shrug*

[Matthew C.]

Well...they're all based on the same basic idea of counting the balls in a player's static zone that he missed and caught and comparing that to a league average for his position...a system I have never been a big fan of...but it could well be that PCA is legitimately missing a tremendous season by Pujols...I jsut never really saw him as that kind of fielder...he was always a good one...but not the first one I thought of in his own league most years. I have a hard time believing stats that show a player having one gigantic career year on defense surrounded by much more modest stuff. *shrug*

Yeah, there are some wild y-t-y fluctuations in all three of those defensive statistics. I guess injuries could be a reason, but I have always wondered how a healthy and young +17 CF one year can be a -12 the next. And for the record, the men invlolved in creating these statistics are aware of these issues and are constantly trying to fox them. I'll give them credit for that.

[SABR Matt]

Oh indeed...I am by no means attacking the guys who create UZR and +/- and the like...what they do is very difficult work and they've put together some systems that generally shine light on a very merky subject. PCA is not without its' major flaws too...it doesn't use full zone data or PBP at the moment...things that it would use now if I had the time and energy to devote to reworking my system. I am merely arguing that the zone based defensive metrics aren't the end of the equation.

[brett]

I don't include offense for pitchers because I don't know the best way to handle the notion that AL pitchers never get to hit. No offensive contribution is not the same as average offensive contribution...or even replacement level. It could wind up biasing the data rather badly.


I think the answer might be to boost the AL league quality as the DH is a method of platooning that allows more players to accrue moderate value.

How was Tulowitski on defense for 2010?

Also, among all time pitchers who do you think tends to get more overrated:
M. Brown (often seen in the 13-18 range); Whitey Ford (often in the mid 20s) or Palmer (late 20s/early 30s)?

[Matthew C.]

I think the answer might be to boost the AL league quality as the DH is a method of platooning that allows more players to accrue moderate value.

How was Tulowitski on defense for 2010?

Also, among all time pitchers who do you think tends to get more overrated:
M. Brown (often seen in the 13-18 range); Whitey Ford (often in the mid 20s) or Palmer (late 20s/early 30s)?

I have seen some leveraging debates lately which show that Ford and Brown were reserved heavily for good teams. Maybe this offsets some of the defensive support issue some. And I showed another defensive-helped guy (Glavine) who faced teams over .500 at the highest clip of any of his HOF contemporaries. Not sure about Palmer. I wonder how much that kind of stuff offsets some of the defensive support. Anyway, I have Ford and Brown in the early 40's and Palmer in the mid- 30's. I have heard a lot of people put Palmer in the early 20's.

[SABR Matt]

Whitey Ford cannot have been facing great OFFENSES...his own TEAM had the great offense.

I think Ford is the most overrated in the group...unlike the other guys on the list, he was not a particularly good K/BB pitcher for a huge chunk of his career...implying that, although he might have been tough to hit, he was not an ideal ace type. DNRA+ sees the last chunk of his career as decided mediocre by ace standards, with him posting a 111 figure. And it does not get better if you look earlier.

[Matthew C.]

Whitey Ford cannot have been facing great OFFENSES...his own TEAM had the great offense.

I think Ford is the most overrated in the group...unlike the other guys on the list, he was not a particularly good K/BB pitcher for a huge chunk of his career...implying that, although he might have been tough to hit, he was not an ideal ace type. DNRA+ sees the last chunk of his career as decided mediocre by ace standards, with him posting a 111 figure. And it does not get better if you look earlier.

Of course he benefitet from not facing the Yankees, but he was witheld for the other good offensive teams often. Well, at least some have suggested. I haven't looked into the facts myself.

[SABR Matt]

What I've seen has only suggested that he faced teams that were above .500 more often (or that the team W% that he faced in total was > .500 by enough to notice). I don't think the offense in isolation has ever been dealt with.

Although...if people are really showing Mordecai Brown in the top 15 all time...then that's a pretty severe overrating of his stature. I have him in the 50s and Whitey in the 90s

[Matthew C.]

What I've seen has only suggested that he faced teams that were above .500 more often (or that the team W% that he faced in total was > .500 by enough to notice). I don't think the offense in isolation has ever been dealt with.

Although...if people are really showing Mordecai Brown in the top 15 all time...then that's a pretty severe overrating of his stature. I have him in the 50s and Whitey in the 90s

And that is just based on PCA, right? Not with all of the advancements of DNRA+ due to lack of PBP data, correct? I know you said Glavine jumped a ton with DNRA+ compared to PCA. Is it possible that something similar would happen to Ford and or Brown if converting PCA to DNRA were possible?

[SABR Matt]

Ford has significant PBP inclusion...so some of his years are DNRA+...but the early years weren't his best, statistically. Anything before 1957 is PCA.

And yes, it's possible that Brown is better than his PCA card looks...if I had that info, he might come up some...I don't think Ford would climb all that much.

[brett]

Ford has significant PBP inclusion...so some of his years are DNRA+...but the early years weren't his best, statistically. Anything before 1957 is PCA.

And yes, it's possible that Brown is better than his PCA card looks...if I had that info, he might come up some...I don't think Ford would climb all that much.

Since Brown is supposed to have had tremendous defenses behind him mightn't he rate even lower in DNRA+?

[SABR Matt]

Since Brown is supposed to have had tremendous defenses behind him mightn't he rate even lower in DNRA+?

He might, yes. No way to know without PBP data.

[RuthMayBond]

He might, yes. No way to know without PBP data.No way to know with it either.

[Matthew C.]

Since Brown is supposed to have had tremendous defenses behind him mightn't he rate even lower in DNRA+?

But PCA already assumes than he had a very good defense.

But DNRA+ includes most every aspect of pitching that guys have some control over...and guys like Brown and Ford, who outperform their peripherals consistanly, tend to do better under those parameters.


We have already established that Glavine and Palmer look a lot better with DNRA+ than PCA and Ford a little better. I bet Brown and a guy like Spahn does too.

[SABR Matt]

No way to know with it either.

Not sure why you would say that...but you most certainly can measure exactly what quantifiable impact a pitcher had on his defense when the ball was in play. It won't ever be PERFECT...but you can get a heck of a lot better idea from DNRA than you can from ERA or even PCA.

[SABR Matt]

Not sure why you would say that...but you most certainly can measure exactly what quantifiable impact a pitcher had on his defense when the ball was in play. It won't ever be PERFECT...but you can get a heck of a lot better idea from DNRA than you can from ERA or even PCA.

Spahn is in the PBP era at least partially...I'll take a look this evening at what I do have for him vs. PCA-BA and see if things are dramatically different.

[AstrosFan]

Matt, can you give us a breakdown of your picks for the Hall on this year's ballot, and how PCA sees them? Really appreciate it.

[RuthMayBond]

please delete