Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Tuesday, September 29, 2009

Rays Index: Debunking The Myth: Wins Is A Useless Statistic For Starting Pitchers

Can we use Wins to evaluate a pitcher over the course of one season? Maybe. We are talking about 28-33 starts. That is still a somewhat small sample size considering the number of factors that are involved. But we can be relatively certain that an 18-game winner is better than a 5-game winner (with similar number of starts). The other variables should be less of a factor in that case. However, when comparing two pitchers with a similar number of wins, those other factors (team defense, scoring, ballpark, etc.) become much more important.

The problem with this post, is that taking a pro-Wins stance leads some to believe that we are anti-other stats. That couldn’t be further from the truth. Stats like ERA+, FIP and tRA are still better measures of how good a pitcher is (although we have minor quibbles with each). However, that does not mean Wins is a useless category. Nor does it mean there are 95 better ways to evaluate a pitcher.

In fact, in the absence of other stats, Wins is a very good, if not great, indicator of a pitcher’s value. So next time you hear somebody say Wins is a crappy way to evaluate a pitcher, throw a drink in their face and then make them read this post.

20 Win tip to Neyer.

Repoz Posted: September 29, 2009 at 04:36 AM | 60 comment(s) Login to Bookmark
  Tags: sabermetrics

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Avoid running at all times.-S. Paige Posted: September 29, 2009 at 04:51 AM (#3335147)
It's certainly a crappy way to distinguish the relative values of CY Young candidates, and that's where I most see the line about wins being a crappy statistic. In any case, it's not surprising that there is a correlation between good pitching and win totals. The exceptions are interesting. Javier Vazquez, for instance, has a career record barely above .500. I wouldn't have guessed that.
   2. The District Attorney Posted: September 29, 2009 at 04:58 AM (#3335151)
in the absence of other stats, Wins is a very good, if not great, indicator of a pitcher’s value.
Fantastic. But there isn't an absence of other stats. Soooo...
   3. Barnaby Jones Posted: September 29, 2009 at 05:06 AM (#3335158)
I think it was Mac over at BravesJournal that said (probably some years ago, at this point) if he could only have one statistic to judge a pitcher's career it would be wins. It's not as silly a statement as I thought on first blush.
   4.   Posted: September 29, 2009 at 05:09 AM (#3335160)
Have I ever mentioned that I hate this kind of argument?
   5. Crashburn Alley Posted: September 29, 2009 at 05:18 AM (#3335170)
In the absence of the Xbox 360, PS3, Wii, original Xbox, Nintendo Gamecube, PS2, PS1, Dreamcast, Nintendo 64, SEGA, Super Nintendo, and NES*, the Atari is a very good, if not great video game system.

* Sega Saturn intentionally left out.
   6.   Posted: September 29, 2009 at 05:20 AM (#3335172)
Boss: Why are you 4 hours late?
Worker: I took the horse and buggy
Boss: What? That's ridiculous
Worker: Au contraire. In absense of other methods of transportations, the horse and buggy is a very good, if not great, way to get to work.
   7. Crashburn Alley Posted: September 29, 2009 at 05:22 AM (#3335173)
I didn't even read the article before making the comment. I like how the author passes off a .5 r-square as a great measure of correlation without testing for statistical significance. It's certainly true that, more often than not, good pitchers will win more games than bad pitchers and conversely, bad pitchers will lose more games than good pitchers. However, what of the glaring middle where up is down and black is white -- Zack Greinke pitching eight shut-out innings to a no-decision, for instance? For a statistic with so much variability, a .5 r-square doesn't impress me any.

And he may have mixed up a correlation coefficient with an r-square, as he talks about a scale of -1 to 1 when it should be 0 to 1 for a squared number (for those of you playing at home, a squared number can't be negative).
   8. bjhanke Posted: September 29, 2009 at 07:10 AM (#3335196)
Of course a squared number can be negative. Just start with an imaginary number, like the number of wins a pitcher (or position player like Pete Rose) contributes with his hustle and clubhouse leadership. Square those claims up and subtract them from the actual wins, just to get the gas out of the colon. - Brock Hanke
   9. Zipperholes Posted: September 29, 2009 at 08:13 AM (#3335200)
This article isn't nearly as silly as it might seem from the blurb, but on the other hand the research isn't very telling either. All he's really showing is Wins correlate well with ERA+ over a four-year period, and thus Wins isn't "useless."

My (and I'd imagine most others') problem with Wins isn't that it's terrible in itself, but it is when used to judge a player on one season, which is the context it's most often used. If it were used only to evaluate several seasons over time, which this study did, it could be a useful stat.
   10. The importance of being Ernest Riles Posted: September 29, 2009 at 12:45 PM (#3335246)
Javier Vazquez, for instance, has a career record barely above .500. I wouldn't have guessed that.

Vazquez is a great example. He's the classic case of a good pitcher on bad teams, and people who must look at wins need to look at them in the context of the teams for which they pitch.

http://www.insidethebook.com/ee/index.php/site/comments/javys_500_record_is_way_above_average/
   11. Best Regards, President of Comfort, Esq. Posted: September 29, 2009 at 12:56 PM (#3335254)
I think it was Mac over at BravesJournal that said (probably some years ago, at this point) if he could only have one statistic to judge a pitcher's career it would be wins. It's not as silly a statement as I thought on first blush.
Sure. If you use any rate stat, they give no context about how long a guy pitched. Over a career, a good pitcher will tend to have a lot of wins. So if you can only judge by one stat, wins is probably the best one.

(Assuming we're excluding complex stats that combine several different stats)
   12. snapper (history's 42nd greatest monster) Posted: September 29, 2009 at 01:00 PM (#3335259)
Sure. If you use any rate stat, they give no context about how long a guy pitched. Over a career, a good pitcher will tend to have a lot of wins. So if you can only judge by one stat, wins is probably the best one.

To add, rate stats give no context on the pitching environment. An ERA of 3.50 looks all shiny, until you realize it's in 1968. Or 7K/9 looks blah, but not in the 1950's. Wins has a lot of context adjustment built in, at least post-1910.
   13. willcarrolldoesnotsuk Posted: September 29, 2009 at 01:01 PM (#3335260)
In fact, in the absence of other stats, Wins is a very good, if not great, indicator of a pitcher’s value. So next time you hear somebody say Wins is a crappy way to evaluate a pitcher, throw a drink in their face and then make them read this post.
In the absence of other stats, so is losses. But not necessarily "more losses are worse than less losses". More losses seems likely to indicate a better pitcher, as he probably got more innings pitched at the major league level than a pitcher with less losses did.

So next time you hear somebody say losses are bad, throw a drink in your face and remember that we don't have an absence of other stats.
   14. Tricky Dick Posted: September 29, 2009 at 01:18 PM (#3335268)
This article defends its thesis by saying that the authors like all of the other stats too. But the article may encourage the many people who choose to view W/L record as superior to any other pitching statistic or whom rely on it to the exclusion of other stats. I get irritated with the many fans and sports writers who judge a starting pitcher as bad or mediocre because they consider the W/L record to be the most significant performance indicator, allowing them to ignore all of the other better statistics which measure performance. The fact that pitcher "wins" can be correllated with performance is not a new finding nor a surprising one. It's a big "so what?" It's not that far removed from finding that pitcher performance is correllated with team W/L record. That wouldn't be surprising, but we would never use team W/L record to evaluate an individual pitcher's performance. By the way, I wonder if "number of no decisions" is correllated with pitching performance measures? I don't know the answer, but if it did, would you use it to evaluate a pitcher?
   15. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 01:48 PM (#3335284)
One statistic I wish someone could come up with, but which I doubt they ever will, is one that can determine which pitchers tend to bring out the best in their teammates, even if it's only for a year or two.

To take a ridiculously extreme example, in 1972 Steve Carlton had an ERA+ of 182, which is great but not really all that historic: It was only the 49th best season in the lively ball era.

And yet that year Carlton was 27 and 10, a pace roughly equivalent to the 1927 Yankees.

But without Carlton, the Phillies were 30 and 87. IOW with any other pitcher they played like the 1962 Mets.

I'm not sure how to explain it. My initial thought was that the Philly regulars weren't really that bad, and that their other starting pitchers were historically dreadful, and hence the contrast. But the 1972 Phillies had a team OPS+ of 82, so so much for that idea. The only metric the Phillies excelled at was that they had the NL's second best fielding percentage, but without a breakdown by pitcher, it's hard to see how that would help us.

But there has to be something more to Carlton's 1972 season than simply randomness. That discrepancy is downright Jekyll and Hyde. If ever there were a case for "intangibles," this would seem to be it.
   16. JPWF13 Posted: September 29, 2009 at 02:25 PM (#3335333)
But we can be relatively certain that an 18-game winner is better than a 5-game winner (with similar number of starts).


Yes, but you can't be relatively certain that the guy with 18 wins is better than the guy with 15.


To take a ridiculously extreme example, in 1972 Steve Carlton had an ERA+ of 182, which is great but not really all that historic: It was only the 49th best season in the lively ball era.

And yet that year Carlton was 27 and 10, a pace roughly equivalent to the 1927 Yankees.


Big deal, .730 winning percentage, you expect a guy with a 182 ERA+ to have a winning % of .746
The Phillies other starters had ERA+s of
84
71
82
89 and
57

The pen wasn't bad, Carlton was awesome...

Ok quick study, 41 Sps since 1901 have had ERA+s between 177 and 187 (90+ ip)
combined w-l: 773-299, .721 winning percentage.

Carlton's expected W-L with an average team behind him in 1972- 27-10
actual record 27-10

so the Phillies likely did play a little better with Carlton pitching- but not remarkably so
If you have a team with an 80 OPS+, it is quite likely that they will hit 95-100 for one pitcher and 60-65 for another...


But there has to be something more to Carlton's 1972 season than simply randomness. That discrepancy is downright Jekyll and Hyde.

??? Carlton WAS GREAT, everyone else sucked rocks, period.

The 72 Phils had a team OPS+ of 82, and ERA+ of 88 (when Carlton is excluded)
The 62 Mets were 82 and 82
the 03 Tigers were 83 and 81

Without Carlton the 72 Phils weren't THAT bad- but they were close.
The year before they went 67-95
The following year 71-91, basically the 71-73 Phillies were a really bad team, 72 being the worst, but by 73 Schmidt and Luzinski were regulars, Garry Maddox and Dave Cash were added the next year...
   17. Greg Pope thinks the Cubs are reeking havoc Posted: September 29, 2009 at 02:38 PM (#3335355)
if he could only have one statistic to judge a pitcher's career it would be wins. It's not as silly a statement as I thought on first blush.

My gut feeling is that innings pitched would be a better stat to judge a career. Just like wins, you'd have to adjust for context, but if a guy's a good pitcher, he's going to get his innings, year after year.
   18. GGC don't think it can get longer than a novella Posted: September 29, 2009 at 02:48 PM (#3335372)
One statistic I wish someone could come up with, but which I doubt they ever will, is one that can determine which pitchers tend to bring out the best in their teammates, even if it's only for a year or two.


Ted Oliver came up with Wins Above Team back in the 1930s. I think Total Baseball listed the leaders in that stat.
   19. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 02:55 PM (#3335379)
Ok quick study, 41 Sps since 1901 have had ERA+s between 177 and 187 (90+ ip)
combined w-l: 773-299, .721 winning percentage.

Carlton's expected W-L with an average team behind him in 1972- 27-10
actual record 27-10


Maybe I'm dense, but where do you get that "an average team" would have given Carlton a 27-10 record? How does that follow from the records of those other 41 starting pitchers? How many of them pitched for "average" teams? Looking down the list, most of them pitched for teams that were well over that, and even the few that were below average (like Greinke's Royals) weren't nearly as bad as the 72 Phillies. Carlton's record in comparison with the rest of his team was so far above it as to be off the charts.

But there has to be something more to Carlton's 1972 season than simply randomness. That discrepancy is downright Jekyll and Hyde.

??? Carlton WAS GREAT, everyone else sucked rocks, period.

The 72 Phils had a team OPS+ of 82, and ERA+ of 88 (when Carlton is excluded)
The 62 Mets were 82 and 82
the 03 Tigers were 83 and 81

Without Carlton the 72 Phils weren't THAT bad- but they were close.


Well, that part I'd already noted. It's the other part that I still have trouble with.
   20. BDC Posted: September 29, 2009 at 02:59 PM (#3335386)
The '72 Phillies scored 3.8 runs per Carlton start, as opposed to 3.0 for the rest of their starters. (Oddly enough, they gave Woody Fryman even better support than Carlton, at 3.9, and Fryman went 4-9 for them as a starter). When neither Carlton nor Fryman was pitching, they scored 2.86 runs per game. The league scored 3.91 that year, so basically with Carlton on the mound they bestirred themselves to become an average offensive team; without him they were terrible.

Carlton himself hit .197 with eight RBI in 1972, not one of his better years at the plate.

And one does wonder if that kind of disproportionate run support (good or bad) can be traced to the pitcher in any way beyond the minor impact of his own bat, or whether it's just the luck of the draw. If Carlton had gotten severely terrible run support, maybe he'd have finished at 22-15 or even 19-18 with the same component numbers, which wouldn't look very historic.
   21. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 03:02 PM (#3335392)
One statistic I wish someone could come up with, but which I doubt they ever will, is one that can determine which pitchers tend to bring out the best in their teammates, even if it's only for a year or two.

Ted Oliver came up with Wins Above Team back in the 1930s. I think Total Baseball listed the leaders in that stat.


Yeah, that was his Kings of the Mound book (or pamphlet), which IIRC dates from the WWII era. It was a great early effort at trying to go beyond raw wins, even if I don't think it accounted for fielding or run support.
   22. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 03:04 PM (#3335395)
The '72 Phillies scored 3.8 runs per Carlton start, as opposed to 3.0 for the rest of their starters. (Oddly enough, they gave Woody Fryman even better support than Carlton, at 3.9, and Fryman went 4-9 for them as a starter). When neither Carlton nor Fryman was pitching, they scored 2.86 runs per game. The league scored 3.91 that year, so basically with Carlton on the mound they bestirred themselves to become an average offensive team; without him they were terrible.

Carlton himself hit .197 with eight RBI in 1972, not one of his better years at the plate.

And one does wonder if that kind of disproportionate run support (good or bad) can be traced to the pitcher in any way beyond the minor impact of his own bat, or whether it's just the luck of the draw.


That's exactly what I was trying to get at in my original post. But I have no idea where the answer lies.
   23. Steve Treder Posted: September 29, 2009 at 03:36 PM (#3335426)
But there has to be something more to Carlton's 1972 season than simply randomness.

No. No, there doesn't. There might be, but by no means does there "have to" be.

Randomness is a completely valid explanation for outcomes of baseball games and all sorts of other phenomena in small samples, and 37 decisions is most definitely a small sample.
   24. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 03:40 PM (#3335431)
But there has to be something more to Carlton's 1972 season than simply randomness.

No. No, there doesn't. There might be, but by no means does there "have to" be.


OK, that was just my artistic side coming out. But as an historian, you might consider digging a bit deeper into that amazing season.
   25. Ron Johnson Posted: September 29, 2009 at 03:47 PM (#3335441)
Andy, why would you think it was anything other than luck? One key point to consider about the Steve inspires the team argument is that at the end of May he was 5-6 with a 2.95 era.

The other thing nobody ever wants to talk about is Carlton's 1973 -- when he seems to have received very poor run support.

A few years back I took a look at what you'd have expected Carlton's record to have been if he's always received team average run support. The answer was 329-238. In other words the guy so frequently cited in the wins carry signal argument actually under-performed over the course of his career (two wins is in the noise of course, but it's two wins the wrong way)

He was +6 over projected in 1971-72, but -2 from 1970-73 so if he had magic it was for a very brief period of time.
   26. BDC Posted: September 29, 2009 at 03:56 PM (#3335453)
the guy so frequently cited in the wins carry signal argument actually under-performed over the course of his career

In the sense that someone who throws 5,200 innings with an ERA+ of 115 can be said to "under-perform," of course.
   27. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 03:57 PM (#3335458)
Ron, I prefaced all this by saying

One statistic I wish someone could come up with, but which I doubt they ever will, is one that can determine which pitchers tend to bring out the best in their teammates, even if it's only for a year or two.


Again, I have no clue as to the cause of all this. But it's not as if there couldn't be factors unique to that team---or to Carlton himself for that one year---that we simply don't know about, or haven't given enough consideration to. I'm skill skeptical about automatically assigning it wholly to randomness.
   28. Steve Treder Posted: September 29, 2009 at 04:14 PM (#3335482)
I'm skill skeptical about automatically assigning it wholly to randomness.

Well, randomness is the null hypothesis. One doesn't accept it, but until a more compelling explanation is identified, it can't be rejected.
   29. zenbitz Posted: September 29, 2009 at 04:22 PM (#3335496)
In the absence of other statistics, a pitchers height is a very good, if not great indicator of a pitchers value.

And Andy - I think the burden is on you to show that the Phillies batting performace under Carlton in 1972 (an increase of 20% or so - not 25% because you have to include their good games with Carlton in their average) is an extreme outlier over 30% of a season (not sure how many no-decisions he had).

But it seems to be just a normal extrema. I should know, I am a Giants fan.
   30. JPWF13 Posted: September 29, 2009 at 04:41 PM (#3335523)
And one does wonder if that kind of disproportionate run support (good or bad) can be traced to the pitcher in any way beyond the minor impact of his own bat, or whether it's just the luck of the draw.

That's exactly what I was trying to get at in my original post. But I have no idea where the answer lies.


It was studied by Bill James in the 1980s, and he determined that a pitcher's run support was essentially random. The team averaged 3.0 per game and gave Carlton 3.8, the next year they scored less for Carlton than they did for other pitchers.

Pick a team this year, any team, I can almost guarantee you that the pitcher with the best run support on that team will have gotten more than 1 run game more than the pitcher on that same team with the worst run support.

Look at the Rays, Garza and Niemann have pitched equally well, but one is 12-6 and one is 8-11, why? Because Niemann has been given 6.07 runs per start and Garza 3.87.
But there is no predictive value to that- Garza is just as likely to have better run support than Niemann next year, as he is to have worse again.
   31. JPWF13 Posted: September 29, 2009 at 04:44 PM (#3335530)
In the sense that someone who throws 5,200 innings with an ERA+ of 115 can be said to "under-perform," of course.


Blyleven is a better example- 4970 IP and a 118 ERA+, but "only" 287-250... not all that "underperformance" is due to poor run support
   32. Jeff K. Posted: September 29, 2009 at 04:47 PM (#3335533)
But it seems to be just a normal extrema. I should know, I am a Giants fan.

Well, so is Steve.
   33. JPWF13 Posted: September 29, 2009 at 04:52 PM (#3335543)
One statistic I wish someone could come up with, but which I doubt they ever will, is one that can determine which pitchers tend to bring out the best in their teammates, even if it's only for a year or two.


Actually now that I think of it, Mike Emeigh actually did compile a list of pitchers by "expected run support"- What the pitcher actually got, versus what the team average was.

If I recall correctly the pitcher who most often "brought" out the best in his teammates (ie they scored more for him than for anyone else*) was Juan Marichal.

*Personally I tend to think Marichal was just lucky in that regard, from 1962-71 Marichal went 202-97 with an ERA+ of 132, a teammate of his went 134-109 with an ERA+ of 119

Marichal routinely received more run support than Gaylord- was it something Marichal DID? Or was he just lucky?
   34. Steve Treder Posted: September 29, 2009 at 04:56 PM (#3335549)
Marichal routinely received more run support than Gaylord- was it something Marichal DID? Or was he just lucky?

Well, see, Marichal was "The Dominican Dandy," and Perry was just "The Rangy Righthander." So there you go.
   35. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 05:01 PM (#3335557)
I'm sure I could go over all 41 of Carlton's box scores for 1972 and it would show that the Phillies scored more runs and made fewer errors behind him than they did behind other pitchers. That still doesn't get us anywhere, because it leaves out the "why?", and 41 games is not that small of a sample size within the context of a season.

Just to cite one possible factor: You often hear fielders tell you how much better it is to play behind a pitcher who pitches quickly; it keeps them on their toes. Obviously that doesn't help a pitcher who keeps giving up home runs, but other things being equal, it would be foolish to dismiss the observation.

And that sort of thing may not be the only motivator. Our knowledge of the psychology of athletes is not yet necessarily at saturation point.

Of course the problem is that it's very tough to control for "other things being equal," so the fallback position becomes "if it can't be measured, it doesn't exist." I guess I just have a problem with that sort of reasoning, which IMO speaks more to a lack of curiosity than to any conclusive theory.
   36. JPWF13 Posted: September 29, 2009 at 05:05 PM (#3335560)
Again, I have no clue as to the cause of all this. But it's not as if there couldn't be factors unique to that team---or to Carlton himself for that one year---that we simply don't know about, or haven't given enough consideration to. I'm skill skeptical about automatically assigning it wholly to randomness.


beating a dead horse some more:
And Andy - I think the burden is on you to show that the Phillies batting performace under Carlton in 1972 (an increase of 20% or so - not 25% because you have to include their good games with Carlton in their average) is an extreme outlier over 30% of a season (not sure how many no-decisions he had).


exactly, a team scoring 20% more for one pitcher than team average is not even an outlier, it happens all the time to virtually every team every year.
   37. BDC Posted: September 29, 2009 at 05:10 PM (#3335565)
The problem with Carlton in this context is that his career was so wildly up and down. The ups were wonderful, but he had seasons where he couldn't seem to buy a win, and others where he was just mediocre. Whatever he had, he kept losing it and refinding it. Now, Whitey Ford, that might be a more interesting story. His teams were pretty strong, granted, but they weren't .690 strong.

Also interesting might be starters who were close to unbeatable in the postseason. Red Ruffing, Lefty Gomez, Bob Gibson, John Smoltz, Curt Schilling. The sample size is tiny but the results are crazy. OTOH it could just be that they were very, very, very good pitchers. Mo Rivera's results are as good in a similar number of postseason innings, and you can't really theorize that a reliever brings out the best in guys who aren't sure in the first seven innings that he'll even appear in the game.
   38. Steve Treder Posted: September 29, 2009 at 05:16 PM (#3335571)
the fallback position becomes "if it can't be measured, it doesn't exist."

No, it doesn't, or at least that isn't what I would conclude, or what the commonly understood scientific method concludes. Not being measured isn't the same thing as not existing, at all.

Moreover, randomness *does* exist. To cite it as a likely explanation of an outcome isn't the same thing as being incurious as to other explanations, rather it's to recognize the real impact randomness has in small sample sizes. And 41 baseball games is, in fact, a small sample size of baseball games, whether within the context of a season (itself not the largest of sample sizes) or not.
   39. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 05:18 PM (#3335573)
And Andy - I think the burden is on you to show that the Phillies batting performace under Carlton in 1972 (an increase of 20% or so - not 25% because you have to include their good games with Carlton in their average) is an extreme outlier over 30% of a season (not sure how many no-decisions he had).


exactly, a team scoring 20% more for one pitcher than team average is not even an outlier, it happens all the time to virtually every team every year.

The undertone of all this chatter is that it's virtually normal for a pitcher to transform a 30-87 team into a 27-10 juggernaut. All it takes is 3.8 RPG, in a year when the NL as a whole was averaging 3.91 RPG. Just imagine what Carlton might have done if he'd had the Astros (4.63 RPG) behind him---he might have gone 35 and 2.

No further explanation needed. Sorry if I'm not signing off on this.
   40. JPWF13 Posted: September 29, 2009 at 05:23 PM (#3335578)
I'm sure I could go over all 41 of Carlton's box scores for 1972 and it would show that the Phillies scored more runs and made fewer errors behind him than they did behind other pitchers.


Well yes they did score more runs for him than for other pitchers (well except for Fryman), but that happens every eyar to every team
a team that scores 4 runs per game doesn't score 4 runs per game for every starter, given 5 starters a typical "spread" would be 3, 3.5, 4, 4.5, 5... but you would also get stuff like 2.75, 3.75, 4, 4.25, 5.25...

Take the 2009 Phillies, 788 runs in 156 games, 5.06 per game.
I added up every 4th game, game 1,5,9,13...and so on, 39 games, 196 runs scored, 5.03 per game
every 2nd 4th game: game 2, 6, 10... and so on 39 games, 178 runs scored, 4.56 r/g
every 3rd 4th game: Game 3, 7, 11... and so on, 198 runs scored, 5.08 per game
every 4th 4th game: agme 4, 8, 12... and so on, 216 runs scored, 5.56 per game

If the Phillies had a set 4 man rotation, their 2nd starter would have gotten 4.56 runs per game, and their 4th starter 5.56...
COMPLETELY RANDOM

There are many things that are remarkable about Carlton's 1972 season, getting .8 more runs per start than team average isn't particularly high up on that list.
   41. Steve Treder Posted: September 29, 2009 at 05:29 PM (#3335583)
There are many things that are remarkable about Carlton's 1972 season, getting .8 more runs per start than team average isn't particularly high up on that list.

Precisely.
   42. JPWF13 Posted: September 29, 2009 at 05:32 PM (#3335592)
The undertone of all this chatter is that it's virtually normal for a pitcher to transform a 30-87 team into a 27-10 juggernaut. A


NO, it is not normal for a 57-97 team to have a pitcher go 27-10, no one said that, Carlton in 1972 was great.

What IS normal is for team to give one starter 5 runs a game and another 4 runs per game AT RANDOM

The Carlton-Phillies and the non-Carlton Phillies were dramatically different teams because Steve Carlton gave up 2.18 runs per 9 innings
and the non-Steve Carlton starters gave up 4.8 runs per 9 innings.

Take the 1966 Dodgers, replace Osteen, Sutton and Drysdale with Fryman, Ken Reynolds and Bill Champion from the 1972 Phillies,
Koufax still goes 27-9, but that team wouldn't even sniff being in contention.
   43. BDC Posted: September 29, 2009 at 05:40 PM (#3335601)
it's virtually normal for a pitcher to transform a 30-87 team into a 27-10 juggernaut. All it takes is 3.8 RPG, in a year when the NL as a whole was averaging 3.91 RPG

Well, no; it takes a pitcher starting 41 games and completing 30, with an ERA+ of 182. The number of pitchers who have done that in a season since the lively-ball era started in 1919? One. Steve Carlton, 1972.

But still, the 27-10 part of Carlton's '72 is tied to run support, obviously.

Edit: shouldn't try to work while posting. Leads to Cokes for JPWF13.
   44. Cork Gaines Posted: September 29, 2009 at 05:48 PM (#3335611)
And yes, in the absence of other stats, Wins still tells a good story.


WAYYYYYYY too much is being made about this one sentence. The point is just that Wins works (to an extent) as a stand-alone stat. I can look at a pitcher’s win total over time and make a general conclusion as to how good a pitcher he is. I would not NEED any other stats.

Now if I want to compare two pitchers with similar win totals, then yes, I will need other stats. But, in general, given enough of a sample size, Wins is a decent stat.

And i am not even suggesting somebody should use Wins. I am just saying it is not as bad of a stat as some would have you believe.
   45. JPWF13 Posted: September 29, 2009 at 06:15 PM (#3335646)
Ok quick study, 41 Sps since 1901 have had ERA+s between 177 and 187 (90+ ip)
combined w-l: 773-299, .721 winning percentage.

Carlton's expected W-L with an average team behind him in 1972- 27-10
actual record 27-10

Maybe I'm dense, but where do you get that "an average team" would have given Carlton a 27-10 record? How does that follow from the records of those other 41 starting pitchers? How many of them pitched for "average" teams? Looking down the list, most of them pitched for teams that were well over that


here is the list:
here

I looked at the team OPS+ for all 41 pitchers, and the average was...95

95, so no, most pitched for teams with poor offenses:
75-79: 2
80-84: 2
85-89: 5
90-94: 10
95-99: 9
100-104: 6
105-109: 3
110-114: 2
which isn't all that unexpected- it is a lot easier to post an era+ of 177-187 with a good dee, and teams that place a premium on good dee tend to sacrifice offense to some extent.

Where Carlton stands out is:
1: His IP; and
2: His rotation mates were unusually bad- none was even an average pitcher.

Te 64 Dodgers were almost as bad at hitting as the 72 Phils, but Koufax had Drysdale- 321 IP 149 ERA+
The 93 Royasl were also a terrible hitting team, but Appier had Cone- hell Pichardo would have towered above the Phils #2 man in 1972....
   46. David Nieporent (now, with children) Posted: September 29, 2009 at 06:19 PM (#3335650)
the fallback position becomes "if it can't be measured, it doesn't exist."
Fair enough; I think unicorns were responsible.
   47. JPWF13 Posted: September 29, 2009 at 06:22 PM (#3335652)
I need a new hobby I think....

But still, the 27-10 part of Carlton's '72 is tied to run support, obviously.


exactly (mathematically I mean)
Carlton gave up 2.18/9, they scored 3.8 for him:
3.8^1.8/(3.8^1.8+2.18^1.8)= .728
.728 * 37 decisions = 26.9 wins, 10.1 losses.

His runs for/runs allowed and his winning percentage were right on the money in 1972.
Given 3.0 runs per game I would expect his W-L would have been: 23.5-13.5
If he'd pitched instead for the 1972 Reds and gotten 4.6/g he likely would have gone: 29-8
   48. Misirlou's been working for the drug squad Posted: September 29, 2009 at 06:46 PM (#3335693)
Zack Greinke - 16-8
Rest of Royals - 48-85

Projected over a full season, the Royals play like the 1986 Mets (108-54). Without him they play like the 1993 Mets (58-104). What's going on?
   49. Walt Davis Posted: September 29, 2009 at 06:48 PM (#3335697)
I can look at a pitcher’s win total over time and make a general conclusion as to how good a pitcher he is.

Wow!! Incredible! You mean you have finally shown (statistically!!) that managers are not complete and utter morons? That they let their best pitchers pitch the most innings and get the most wins (and losses)?

You, sir, are a bright light in our sabermetric darkness.

Now, can anyone explain the following conundrum to me: if singles hitters are so crappy, why are 29 of the top 30 on the career singles list in the HoF?
   50. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 07:08 PM (#3335724)
I guess that your (meaning everyone's but mine) bottom line is that in 1972 Carlton had a fabulous ERA+ / number of starts / number of complete games, and with 3.8 RPG it was almost to be expected that he would go 27 and 10.

You note that this extra run support for Carlton is relatively common for pitchers, and that it happens all the time.

Fine. But then beyond attributing that extra run support to mere randomness, wouldn't it first make sense to see who Carlton's mound opponents were, and how the Phillies scored against them, compared to how other teams did?

If Carlton was the fortunate beneficiary of being matched up against stiffs, then that's one thing. If that were the case, then that extra run support would have been expected.

But if the Phillies consistently hit better against some of the league's better pitchers than would have been expected, given the matchups, then that's another thing.

I won't jump the gun on this, but I'd be very surprised if Carlton didn't face a disproportionate number of aces and number 2 starters, and relatively few stiffs. Which would make that 27 and 10 record even more unlikely and extraordinary, and IMO less likely to be explained by randomness.

One "non-random" explanation might be the Phillies were motivated by the buzz surrounding Carlton's extraordinary year, and performed above their true speed. Which in fact they did, as attested to the extra run support that Carlton received.

The other possibility is that Carlton---perhaps wanting to show up the Cardinals for making him the victim of a spite trade to a last place team due to a salary dispute, and knowing the caliber of his opposing pitcher---"pitched to the score."

I know what that cliche generally means: a Jack Morris who gets staked to six runs and then gives five runs back before miraculously bracing himself and holding on for the win. But that's not what I'm referring to.

I'm talking about a great pitcher, who knows he's facing a superior team with one of their best pitchers. And who knows that he's going to have to bear down from the first pitch in order to have a chance of winning. IOW in this case, "pitching to the score" meant thinking he'd need a shutout in order to win.

We all agree that the Phillies played like a much better team behind Carlton---instead of being the worst scoring team in the NL, they were only slightly below average. What I'm trying to figure out, beyond Carlton's numbers alone, is why. Why did they play so much better?

And if you still want to say it's all random good fortune, I want to ask you one more question: What role, if any, do you think that human psychology or emotion might have played in any of this? Do you think it's impossible that the Phillies simply felt more like a Major League team with Lefty out there giving them a chance to win every time they took the field? As opposed to what they must have felt like with those minor league level pitchers they had out there nearly every other day?

Funny, but IIRC I read more than a few comments to that effect from the Phillies during that 1972 season. I guess I should have ignored them, but then those comments seemed to ring true more than not.

Steve, you wrote this:

Moreover, randomness *does* exist. To cite it as a likely explanation of an outcome isn't the same thing as being incurious as to other explanations,


Well, what "other explanations" might enter your mind? Where might your own curiosity lead you, beyond your mathematical instincts?
   51. Steve Treder Posted: September 29, 2009 at 07:30 PM (#3335772)
Well, what "other explanations" might enter your mind? Where might your own curiosity lead you, beyond your mathematical instincts?

The psychological stuff you talk about is entirely plausible. Baseball players aren't machines; they're humans as prone to emotional ups and downs and motivational/confidence streaks and slumps, just like all the rest of us.

But here's the larger issue with Carlton and the 1972 Phillies: as far as I've ever seen, there is no *necessity* for anything other than the dull, boring cause of randomness to entirely explain the performance of the team in Carlton's starts. That sort of swing in run production between starters, or something close to it, happens *all the time.* We became very aware of it in Carlton's season because of the spectacular nature of the 32-87 vs. 27-10 comparison, but that comparison has vastly more to do with how crappy the rest of the Phillies' pitching staff was (and how brilliant Carlton was) than the run production anomaly.

So, given that the run production anomaly is the sort of thing that occurs on lots of teams in many, many seasons, I hope you can understand why I'm less than strongly compelled to look closely at the 1972 Phillies. There are many more puzzling, engaging questions out there.
   52. BDC Posted: September 29, 2009 at 07:31 PM (#3335776)
I can look at a pitcher’s win total over time and make a general conclusion as to how good a pitcher he is

Not to pile on your inaugural post too much, Cork Gaines, but the problem is that even with large sample sizes, even Career Wins is not much use. I mean, Phil Niekro and Pedro Martinez have both had long careers, and Niekro has half-again as many Wins. That tells you almost nothing in terms of comparing them. Don Sutton had 70 wins more than either Carl Hubbell or Bob Gibson, in long careers. This is not a useful thing to know, because the latter two were highly dominant pitchers in their day, MVPs of their leagues, and Sutton was just plain never that good.
   53. BDC Posted: September 29, 2009 at 07:35 PM (#3335783)
One "non-random" explanation might be the Phillies were motivated by the buzz surrounding Carlton's extraordinary year, and performed above their true speed. Which in fact they did

But Andy, that still leaves you with the Woody Fryman factor. Woody had a certain je ne sais quoi, but I don't see how it translated into an extra run per game of support.
   54. JPWF13 Posted: September 29, 2009 at 07:40 PM (#3335796)
The other possibility is that Carlton---perhaps wanting to show up the Cardinals for making him the victim of a spite trade to a last place team due to a salary dispute, and knowing the caliber of his opposing pitcher---"pitched to the score."


Doubtful, his pythag WP and his actual WP match perfectly.

I won't jump the gun on this, but I'd be very surprised if Carlton didn't face a disproportionate number of aces and number 2 starters, and relatively few stiffs. Which would make that 27 and 10 record even more unlikely and extraordinary, and IMO less likely to be explained by randomness.


since you ask:
after 11 starts he was 5-6, the average ERA+ of his opposing SP? 116.
then he had 9 straight starts where his average opponent was an 86...
11 straight where the average was 111 (no one was average, he was either facing guys over 130 or under 90...)
Then his last 10 averaged 92.
All told, his average Opponent had an ERA+ of 103,
14 were under 90
7 in the 90s
9 100-120
11 over 120

All told his opponents seemed to tended to be of the stars or scrubs variety
   55. JPWF13 Posted: September 29, 2009 at 08:04 PM (#3335849)
Fine. But then beyond attributing that extra run support to mere randomness, wouldn't it first make sense to see who Carlton's mound opponents were, and how the Phillies scored against them, compared to how other teams did?


Look, there is a point, where yes, I'd be curious, if a team averages 3 overall, but 6 for one guy, something that extreme (given Carlton's # of starts maybe 5r/g would be where I'd go, "wow, what is going on".

Carlton tended to have a personal catcher, McCarver, maybe he had a good year (just checked, not really)...

Carlton was great in 1972, but he started out 5-6, at what point did his team decide, hey this a special year, let's start playing better? Perhaps it was after those first 11 starts when he mowed through the 86 ERA+ riff raff in his next 9 starts?
and why did they go in the tank the next year?

IP 300+, ERA+ 170+ (1920 to date:
Cnt Player            ERA+  W  L   IP  Year Age Tm
+----+-----------------+----+--+--+-----+----+---+---+
    
1 Bob Gibson         258 22  9 304.2 1968  32 STL 
    2 Dolf Luque         201 27  8 322   1923  32 CIN 
    3 Hal Newhouser      195 25  9 313.1 1945  24 DET 
    4 Carl Hubbell       193 23 12 308.2 1933  30 NYG 
    5 Sandy Koufax       190 27  9 323   1966  30 LAD 
    6 Wilbur Wood        188 22 13 334   1971  29 CHW 
    7 Vida Blue          185 24  8 312   1971  21 OAK 
    8 Steve Carlton      182 27 10 346.1 1972  27 PHI 
    9 Dazzy Vance        174 28  6 308.1 1924  33 BRO 
   10 Thornton Lee       173 22 11 300.1 1941  34 CHW 
   11 Red Faber          171 25 15 330.2 1921  32 CHW 
   12 Gaylord Perry      170 24 16 342.2 1972  33 CLE 

nice company
BTW that 1972 Cleveland team hit just as badly as the 1972 Phillies...
they went 72-82, despite having an above average rotation (110 ERA+)
The Indians scored 2.95 runs per Start for Gaylord, their average was 3.03...
He went 24-16, versus Carlton's 27-10...
   56. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: September 29, 2009 at 08:46 PM (#3335943)
Well, what "other explanations" might enter your mind? Where might your own curiosity lead you, beyond your mathematical instincts?

The psychological stuff you talk about is entirely plausible.


Which is all I'm getting at to begin with. I'm not denying the possibility of randomness, but only saying that it's not something that you can "know" on the same level that you can "know" the numbers on BB-ref. To me it's just not a closed case.

Carlton was great in 1972, but he started out 5-6, at what point did his team decide, hey this a special year, let's start playing better? Perhaps it was after those first 11 starts when he mowed through the 86 ERA+ riff raff in his next 9 starts?

Could be. At some point his teammates were bound to catch on that this was one special pitcher, and that kind of a streak can lend credibility to that thought.

and why did they go in the tank the next year?

Different team, different pitcher, different psychology and different set of circumstances. 1972 was obviously his career year, and the next year he got into a funk with the reporters when they discovered his unorthodox training regimen, and wound up as a below average pitcher.
   57. Steve Treder Posted: September 29, 2009 at 08:57 PM (#3335957)
I'm not denying the possibility of randomness, but only saying that it's not something that you can "know" on the same level that you can "know" the numbers on BB-ref. To me it's just not a closed case.

I'm pretty sure nobody's saying it is. But the issue is that the distribution of run production exhibited by the 1972 Phillies is well within the range we see all the time. It's not nearly unusual enough to require any explanation beyond random fluctuation, and so doesn't call out for any particularly close inquiry.

So, to make the case that the 1972 Phillies displayed exceptionally unusual psychological dynamics, then you'll have to provide far more evidence than their run production pattern behind Carlton. That piece of evidence doesn't survive the first random-error cut.
   58. bads85 Posted: September 29, 2009 at 10:29 PM (#3336033)
One thing that hasn't been brought up yet about win totals for a starter in this discussion is bullpen support -- something that Carlton rarely needed that year. The Phillies' pen was atrocious that year, but because Carlton went so deep in so many games, the pen didn't have too many opportunities to blow his wins. The pen only used in eleven of his forty-one appearances and only blew one lead. In Carlton's four no decisions, the pen entered when he was already behind.

Carlton himself limited some of the "badness" of the Phillies by making sure they didn't come in the game on most of the days he pitched.
   59. Howie Menckel Posted: September 29, 2009 at 11:41 PM (#3336065)
I wonder if bad teams with great pitchers tend to send out their best players for those games, and rest the regulars only when other pitchers are going.
One could make the argument to go the other way - maybe Carlton can win with anyone, but we need to stack the lineup for bums - but I'd be surprised if that happens.

Anyway, his 27-10 doesn't seem that weird to me given his utter dominance, either. The 30 CG is key - harder to win 27 when some lesser RPs are eating IP whenever you struggle a little.
   60. Howie Menckel Posted: September 29, 2009 at 11:41 PM (#3336066)
guh, half a Coke to bads85

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Adam S
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogTigers' Miguel Cabrera appears to re-injure ankle, leaves game
(9 - 3:35am, Sep 01)
Last: Cooper Nielson

NewsblogBackman named PCL’s top manager
(12 - 2:57am, Sep 01)
Last: ReggieThomasLives

NewsblogBob Melvin calls Athletics 'pathetic' after Angels sweep four-game set
(1 - 2:09am, Sep 01)
Last: Walks Clog Up the Bases

NewsblogOT: NBC.news: Valve isn’t making one gaming console, but multiple ‘Steam machines’
(779 - 2:09am, Sep 01)
Last: DJS and the Infinite Sadness

NewsblogJesus Montero gets heckled by Mariners cross checker during rehab stint
(58 - 2:00am, Sep 01)
Last: Weratych

NewsblogAthletics Acquire Adam Dunn
(22 - 1:37am, Sep 01)
Last: 6 - 4 - 3

NewsblogOMNICHATTER 9-1-2014
(1 - 1:13am, Sep 01)
Last: Best Regards, President of Comfort, Esq.

NewsblogOT August 2014:  Wrassle Mania I
(90 - 12:53am, Sep 01)
Last: andrewberg

NewsblogOT: Politics, August 2014: DNC criticizes Christie’s economic record with baseball video
(6296 - 12:32am, Sep 01)
Last: David Nieporent (now, with children)

NewsblogOMNICHATTER 8-31-2014
(100 - 12:29am, Sep 01)
Last: Jose Can Still Seabiscuit

NewsblogHigh School Baseball Game In Japan Takes 50 Innings, Four Days
(6 - 11:46pm, Aug 31)
Last: Gamingboy

NewsblogBlue Jays Acquire Mayberry Jr.
(2 - 11:41pm, Aug 31)
Last: John Northey

NewsblogPhoto of the day: Bill Murray, indy league ticket-taker
(42 - 11:33pm, Aug 31)
Last: Robert in Manhattan Beach

NewsblogSherman: How Reds react to second-half swoon will be major factor in offseason
(5 - 9:46pm, Aug 31)
Last: Select Storage Device

NewsblogOrioles Acquire Kelly Johnson
(20 - 9:44pm, Aug 31)
Last: Mike Emeigh

Page rendered in 0.9024 seconds
52 querie(s) executed