PCA Request Line

[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]

Originally Posted by 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?

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

[brett]

Originally Posted by STLCards2

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]

Originally Posted by 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?

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. [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]

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.

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. [quoteOf 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.[/quote]

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

6.

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.

[quoteAnd if you folks want to dismiss me because I have a harsh tone in this thread...fine, whatever...I get grumpy when confronted with a scientific fraud.[/quote]

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.

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]

Originally Posted by leewileyfan

[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]

Originally Posted by brett

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]

Originally Posted by 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.

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.

[STLCards2]

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.

[STLCards2]

Originally Posted by 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.

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.

[STLCards2]

Originally Posted by 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.

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]

Originally Posted by 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?

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]

Originally Posted by SABR Matt

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]

Originally Posted by SABR Matt

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]

Originally Posted by brett

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.

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]

Originally Posted by STLCards2

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]