The most important pitching stat

[Bill Burgess]

What Bill wants to do I think is create a win-loss metric...or a winning percentage like statistic to make it easier to see and comprehend the context in which a pitcher pitched.

He wants an offense-independent Runs Scored For and a Defense Independent Runs Allowed By...so he can put them together and see the pitcher's true W-L

That would be an interesting way to view pitching skill.

Bill doesn't have a clue what he wants, but he trusts Matt does. Going back to the Matty/Brown example, we need to adjust for Brown having the better team.

I don't know if the Cubs were better offensively, or better defensively. But I must believe that a way can be discovered by Matt to make these adjustments.

Bill

[Sultan_1895-1948]

You need a "necessary runs" formula that will include team defensive rank, park factors, and how the planets were aligned Should be simple enough.

[RuthMayBond]

Obviously, W-L % is the single most important stat.
Otherwise, Bert Blyleven would be elected to the Hall of Fame some time ago.

If only that explained Ryan, Eck, Lyons, Niekro, Roberts . . .

[SABR Steve]

1. ERA
2. Innings
3. ball park
4. Completions
5. total run average
6. total bases allowed per game
7. wins above .500

[SABR Steve]

Your forgetting that Martinez needs relief, Johnson didn't. What would Martinez' ERA look like had he been forced to go nine? To me it's a combination of effectiveness and endurance.

[Dasperp]

Your forgetting that Martinez needs relief, Johnson didn't. What would Martinez' ERA look like had he been forced to go nine? To me it's a combination of effectiveness and endurance

And what would Johnson be like if he had to pitch off a lower mound? or in today's smaller ballparks? or against today's weight training (and steroid using) sluggers? or against black and latino players? There are so many problems you encounter when comparing players of such different eras that you can't just say Pedro couldn't have gone nine innings in a four man rotation and have that be the end of it.

[BoSox Rule]

Your forgetting that Martinez needs relief, Johnson didn't. What would Martinez' ERA look like had he been forced to go nine? To me it's a combination of effectiveness and endurance.

This is such an original user name

[digglahhh]

And what would Johnson be like if he had to pitch off a lower mound? or in today's smaller ballparks? or against today's weight training (and steroid using) sluggers? or against black and latino players? There are so many problems you encounter when comparing players of such different eras that you can't just say Pedro couldn't have gone nine innings in a four man rotation and have that be the end of it.

Agreed.

Some assumptions seem more abled to be extrapolated than others (increasing IP vs. hypothesizing the impact of segregation), but that doesn't make the conclusions any more factual.

I think this is a problem. Numbers are better measuring some things than others (the more tangible the better) and therefore we feel that we can make such assumptions with confidence. Conversely, the abstract considerations are not easily hypothesized, so they are treated as unaddressable. This is the case with the respective concerns for Johnson and Martinez.


BTW, I don't think its a foregone conclusion that Pedro pitching more innings wouldn't have made his season better. By extending Pedro to 275 innigs at say, 85% of his relative effectiveness, you might produce what is considered an even greater result once sample size is considered. In effect, you would be minusing the only knock one can have against that season. I am certainly unconvinced you would wind up with enough of a drop off in overall dominance to actually make it a "worse" season.

Just for kicks let's take this hypothetical:

We'll extend Pedro's IP by 1/3 to 289

then we'll assume that over those extra 72 innings he pitched to the tune of a 3.00 ERA (a 58% increase in ERA)

Pedro would then finish that season with a 2.05 ERA over 289 innings as opposed to a 1.74 ERA over 217. Which is more impressive? I guess that depends on your taste

Assuming one K per inning over this streach (his actual K/9 was almost 12, that season) he would also wind up with 356 strikeouts.

This is not to mention how his win, CG or shutout totals may also inflate.

His Rel. ERA undajusted for ballpark would still 241, and he would have led the league in IP by over 50.

The point is, even extrapolating a severe drop-off in efficiency and compounding Pedro's season, you get one of utter (possibly even greater) domination.

All these adjustments are abitrary, but the point remains rather clear. No adjustment of that season for workload, etc. can mute how amazing that season was.

[SABR Matt]

Agreed.

Some assumptions seem more abled to be extrapolated than others (increasing IP vs. hypothesizing the impact of segregation), but that doesn't make the conclusions any more factual.

I think this is a problem. Numbers are better measuring some things than others (the more tangible the better) and therefore we feel that we can make such assumptions with confidence. Conversely, the abstract considerations are not easily hypothesized, so they are treated as unaddressable. This is the case with the respective concerns for Johnson and Martinez.


BTW, I don't think its a foregone conclusion that Pedro pitching more innings wouldn't have made his season better. By extending Pedro to 275 innigs at say, 85% of his relative effectiveness, you might produce what is considered an even greater result once sample size is considered. In effect, you would be minusing the only knock one can have against that season. I am certainly unconvinced you would wind up with enough of a drop off in overall dominance to actually make it a "worse" season.

Just for kicks let's take this hypothetical:

We'll extend Pedro's IP by 1/3 to 289

then we'll assume that over those extra 72 innings he pitched to the tune of a 3.00 ERA (a 58% increase in ERA)

Pedro would then finish that season with a 2.05 ERA over 289 innings as opposed to a 1.74 ERA over 217. Which is more impressive? I guess that depends on your taste

Assuming one K per inning over this streach (his actual K/9 was almost 12, that season) he would also wind up with 356 strikeouts.

This is not to mention how his win, CG or shutout totals may also inflate.

His Rel. ERA undajusted for ballpark would still 241, and he would have led the league in IP by over 50.

The point is, even extrapolating a severe drop-off in efficiency and compounding Pedro's season, you get one of utter (possibly even greater) domination.

All these adjustments are abitrary, but the point remains rather clear. No adjustment of that season for workload, etc. can mute how amazing that season was.

Nice post digglahhh.

I do think however that we can get some idea of how a pitcher would respond to an increased workload by studying that pitcher's effectiveness as his innings mount in each game and as they mount over the course of a full season. That's one form of situational analysis I intend to explore at some point down the road.

[digglahhh]

Thanks Matt.

I would think your idea is possible, and I'd be interested in seeing what you find,but there are so many variables, opposing team, situation...

I would need to see a crystal clear trend to make a determination as to a direct causal relationship. The idea that as a pitcher pitches more innings in more games, and then more games, he weakens is certainly logistically syllogistic, but I'm not going take it as a given.

I don't know if we really clash as much as it seems. I just think that I have a high burden of proof when evaluating your type of work. I'm willing to allow a lot more random variation in numbers than most, because I feel the game has infinite variables and situations.

As we've seen with Pedro, the accomplishments over 217 innings leave plenty of room for a drop off while still maintaining "dominance" throughout the season, at least on paper.

If it becomes clear that there is a directly proportionate relationship between Pedro's attrition and number on innings, or pitches, or games throughout a game/season. I'd have to recognize it.

The nature of the downturn would be important too. If it was a deliberate but relatively slowly incremental process I wouldn't mind so much, b/c at this rate he still would probably be better than any choice in a manager's bullpen, save perhaps the closer. If the curve was exponential, then that would show that there was clearly a wall to Pedro's effectiveness.

One thing that has always amazed me about Pedro is how he can strike out so many but still go deep into games with a relatively low pitch count. Pedro as dominant a K artist as anyone, but what makes him unique is that his make-up is more like that of a Greg Maddux than a Randy Johnson.

[Brew Fan]

Winning % above team

http://baseballgm.blogspot.com

[csh19792001]

Winning % above team

http://baseballgm.blogspot.com

That's actually a very flawed way of measuring things (I once thought it was a very accurate way of assessing things, as well). Why?

Pretty simply... a guy on a great team can't seperate himself as much in terms of percentage as a guy on a bad team.

Example: Even if Maddux had gone 21-0 in 1995, he wouldn't have been able to even come close to surpassing Steve Carlton's .461 pt WPCT differential from his 1972 campaign. Totally unfair to penalize Greg further due to being on a great team.

There's an excellent article in the latest SABR Baseball Research Journal (from which I excerpted the example above) that looks at normalized winning percentage used in combination with Wins Above Team (Palmer's Total Baseball metric). I think it's more fair, and doesn't double penalize pitchers who were, overall, on above average teams.

[SABR Matt]

It still misses another important point. Pitchers on the same team don't necessarily get the same run support. You don't believe me...you talk to Ryan Franklin.

[csh19792001]

It still misses another important point. Pitchers on the same team don't necessarily get the same run support. You don't believe me...you talk to Ryan Franklin.

Oh, I know. There's definately more to the story, you're absolutely right.

Ever heard of the Marichal-Perry syndrome? You can google it. A blogger came up with that and MOWP (Median Offensive Winning Percentage).

The link is dead, which SUCKS. But I found some fantastic stuff there. Very equilizing.

http://baseballblogs.org/entries.php...ate=2004-08-16

http://runsupportindex.blogspot.com/...x_archive.html

[Dontworry]

Both ERA and WHIP are flawed, and the flaws in both are related to the defense behind the pitcher.
A pitcher for a strong defensive team (one that takes away hits with its defense) will have a better WHIP than the same pitcher for a weak defensive team (one that gives away hits that might otherwise be outs).
On the other hand, a pitcher for an error-prone team will see his ERA benefit from those errors when they're followed by hits and unearned runs, even though he's the one letting up those post-error hits. The same pitcher for a team that doesn't make errors will see his ERA take a hit because of the lack of unearned runs.
Give me RPG -- runs allowed per game (nine innings).