Play Index Season Finders Now Sort by WPA Clutch
Posted by Sean Forman on July 15, 2010
I've added WPA Clutch (which is the context-rich performance minus the context-neutral performance) to the season finders. Here are the top and bottom five hitters in Clutch. Full Description of Win Probability Added including Clutch.
| Rk | Player | Clutch | From | To | PA | Tm |
|---|---|---|---|---|---|---|
| 1 | Nellie Fox | 13.4 | 1950 | 1965 | 10035 | CHW-HOU |
| 2 | Tony Taylor | 10.5 | 1958 | 1976 | 8501 | CHC-TOT-PHI-DET |
| 3 | Tony Gwynn | 10.0 | 1982 | 2001 | 10232 | SDP |
| 4 | Pete Rose | 9.3 | 1963 | 1986 | 15861 | CIN-PHI-TOT |
| 5 | Sandy Alomar | 8.0 | 1964 | 1978 | 5160 | MLN-ATL-TOT-CHW-CAL-NYY-TEX |
| Rk | Player | Clutch | From | To | PA | Tm |
|---|---|---|---|---|---|---|
| 1 | Sammy Sosa | -16.8 | 1989 | 2007 | 9896 | TOT-CHW-CHC-BAL-TEX |
| 2 | Frank Robinson | -15.9 | 1956 | 1976 | 11743 | CIN-BAL-LAD-CAL-TOT-CLE |
| 3 | Jim Thome | -13.0 | 1991 | 2010 | 9638 | CLE-PHI-CHW-TOT-MIN |
| 4 | Lance Parrish | -12.7 | 1977 | 1995 | 7797 | DET-PHI-CAL-TOT-CLE-PIT-TOR |
| 5 | Mike Schmidt | -10.6 | 1972 | 1989 | 10062 | PHI |
I'm going to let someone else tell Frank Robinson about his ranking here.
July 15th, 2010 at 11:43 am
Reading WPA, I have a quick question:
Is score only considered in terms of scoring differential, or the actual score of a game? What I mean to say, is a 9-8 game considered the same as a 1-0 game, all other things (inning, base-out) being the same? Because, intuitively at least (and I think the guys over at Football Reference investigated it), it seems that it is easier to overcome a given deficit in a high scoring game than in a low scoring game. I realize football works different than baseball, and I suppose the counter would be that it's also easier for the deficit to grow, but I'm curious how this is considered.
Thanks!
July 15th, 2010 at 11:48 am
It's only by margin, but the values are adjusted by the run scoring environment, so a hitters park will have different values from a pitcher's park.
July 15th, 2010 at 11:55 am
Gotcha. Makes sense. Is it also adjusted for the actual quality of the players? For instance, a 1 run deficit with no outs in the 9th against Rivera is very different than the same situation against a far inferior pitcher. Or, at least it would intuitively seem that the chances of winning the former situation are less than winning the latter. Or am I misunderstanding?
July 15th, 2010 at 11:56 am
No it compares the player to average player, so while the Royals have lower prob coming back against Mo than they do Aardsma it treats them the same.
July 15th, 2010 at 11:57 am
Okay. Does that have any potential outcome on the data? Or does that stuff A) still not matter THAT much in the grand scheme, B) tend to average out, or C) both?
July 15th, 2010 at 11:58 am
I would assume it would even out, but I'm sure we could find examples where a batter had been dealt a really bad hand.
July 15th, 2010 at 1:40 pm
Keep in mind that every stat is always a compromise betwen completeness and excluding those things that are expected to more or less even out over time. Including every aspect of context would mean you are not looking at a stat any morer, you're watching the actual game. Assuming that the quality of the players in differing level of clutch situations will more or less even out over hundreds or thousands of PAs is probably a pretty good, useful assumption, but of course not perfect.
Sean, you continue to do an extraordinary job taking intersting new stats and incorporating them into your brilliantly useful and accessible system of searching and organizing. Just amazing!
July 15th, 2010 at 4:54 pm
Birtelcom-
You're absolutely right. I hope my questioning didn't come off as somehow doubting the stat. I was just curious to understand how it worked. As much as I try to understand the new stats, I am still somewhat of a layman, so I ask questions. Sean answered each one, which not only clarified it to me, but also demonstrated the completeness of the thought that went into it. Thanks!
July 16th, 2010 at 12:14 pm
Fantastic job on all these stats. BB-Reference has become my #1 spot for baseball.
I also posted this at Tango's site, as I am having a little difficulty with this:
Nellie Fox's difference in clutch is due to his batting average, basically hitting singles. He hit a few more triples, but less doubles, virtually same homeruns and walks.
In EVERY year, 1950-1964, his batting average was worse in Late & Close versus High Leverage. For his career, he was .051 worse.
In every year, his differential was more than the league differential - the odds should be 32,767 to 1 against that.
In 1950, his L&C ba was .171; his high leverage .362
I would guess that there is some amount of overlap in the 2 categories.
I would like to reconcile the two. I don't know if anyone has the data for one of his years, preferably 1959 where there is also a big gap (and a Sox pennant).
It just appears to be not right. I would like to prove myself wrong.
Thanks
July 17th, 2010 at 12:22 pm
Good question Dquinn. I was looking for the definition of "High Leverage" and I don't see it, does anyone know? He would have a bit over 100 PA a season of both Hi-Lev and L&C PA and I agree it seems likely there would be a lot of overlap.
It looks like Fox was a good situational hitter, particularly high percentages of productive outs, and those could help his WPA in certain situations while hurting his BA. Still doesn't answer your question however.