Comments on: David Freese: now THAT was the best World Series performance in history This and that about baseball stats. Tue, 16 Jul 2013 17:01:55 +0000 hourly 1 By: scott-53 Sun, 30 Oct 2011 01:44:02 +0000 @ 91: Not sure.

By: Richard Chester Sun, 30 Oct 2011 00:27:20 +0000 @94

In the 1976 ALCS deciding game Chris Chambliss led off the ninth inning with a walkoff HR and got a WPA of .37. He hit the first pitch. In the first game off the1949 WS Tommy Henrich's lead-off walk-off HR got him .36, I don't remember which pitch he hit.

By: Andy Sun, 30 Oct 2011 00:24:10 +0000 I suspect that it's due to different run-scoring environments. Run expectancy is of course different when overall scoring is higher or lower, and WPA can be calculated based on any run-scoring environment you wish. I don't know how B-R's WPA values are calculated but I'm willing to bet that Sean taken either league yearly run scoring and/or park environment into account, so there are slight differences.

Just to explain it a bit more, if you think back to Coors Field in the late 1990s, a 2-run lead in the 9th was less safe than it is there today, since so many more homers were hit there back then, so the WPA for the 1990s Coors Field scenario would predict a greater likelihood of the home team winning when going to the bottom of the 9th down by 2 runs than it would in the 2011 Coors Field scenario.

Quite often, WPA simply uses the entire database of baseball history when running scenarios--this ignores run-scoring environment and is fine as an average. One can choose to take a subset based on league-wide scoring or park, which might make it more accurate, but of course also draws on a smaller set of historical games.

By: Richard Chester Sun, 30 Oct 2011 00:06:07 +0000 @90, @91

I would believe that the WPA for those game-winning HRs is solely dependent on the wWE at the end of the top of the ninth inning. Sometmes it is 63% and sometimes it is 62%. Don't ask me why.

By: Mike Felber Sat, 29 Oct 2011 20:46:01 +0000 Vidor, I also have differed on the ambiguity in, or conflation of, highest & best re: WPA. But it is not right to call names & be nasty when critiquing an idea. If you would not do it in person-or to someone bigger 'n you-it is not right & certainly not brave to do it here.

Why don't you just apologize now & thus likely a avoid a probable ban?

By: Andy Sat, 29 Oct 2011 19:03:29 +0000 My understanding is that some versions of WPA use ball/strike count (for example ESPN Game Cast) but B-R's does not. I have never heard of any version that uses order position, although this is of course a major failing of WPA.

Vidor, enjoy the weekend posting because I suspect the bosses will ban you when they see your comments on Monday. If you can't criticize without using that sort of language, they won't let you contribute here. Incidentally I have nothing at all to do with such decisions.

By: John Autin Sat, 29 Oct 2011 18:54:53 +0000 @89-90 -- Are you 100% sure that WPA encompasses the ball/strike count and/or the batting order?

I did not think it covered either of those things. I can't find a definitive source stating what WPA doesn't cover, but the basic descriptions on B-R and on FanGraphs both omit any mention of count or batting order. If there's a source that says count & order are included in WPA, I hope you'll point me to it.

I would have thought that the different values mentioned @88/90 derived from different offensive contexts.

By: Doug Sat, 29 Oct 2011 17:41:08 +0000 @89.

Thanks Scott, I think you've got it.

I also looked up Fisk's HR in 1975 game 6, which was also leading off the inning. He batted fourth and hit a 1-0 pitch (same as Mazeroski) for a 0.36 WPA.

By: scott-53 Sat, 29 Oct 2011 11:23:09 +0000 @88: I'm thinking it was the ball/strike count.

1-0 against Mazeroski the #8 hitter.

3-2 against Freese the #6 hitter

By: Doug Sat, 29 Oct 2011 06:21:34 +0000 I looked up the WPA just for the one play of Freese's game-winning HR. It was 0.38.

Then I looked up WPA for Mazeroski's game-winning HR in game 7 in 1960. It was 0.37.

I know they're real close, but shouldn't they be identical? Tie score, leading off bottom of 9th or later inning. Win probability-wise, it has to be an identical sitiuation, doesn't it?