THE BOOK--Playing The Percentages In Baseball

Sunday, July 11, 2010

Text of a new counting stat + talk of WAR

Poz points out something he read on some website, about some new counting stat:

The pitcher whose team assumes a lead while such pitcher is in the game, or during the inning on offense in which such pitcher is removed from the game, and does not relinquish such lead. That is unless: (1) such pitcher is a starting pitcher and Rule 10.17(b) applies; or (2) Rule 10.17(c) applies. 10.17(b) If the pitcher whose team assumes a lead while such pitcher is in the game, or during the inning on offense in which such pitcher is removed from the game, and does not relinquish such lead, is a starting pitcher who has not completed (1) five innings of a game that lasts six or more innings on defense, or (2) four innings of a game that lasts five innings on defense, then the official scorer shall credit as the winning pitcher the relief pitcher, if there is only one relief pitcher, or the relief pitcher who, in the official scorer’s judgment was the most effective, if there is more than one relief pitcher. (c) The official scorer shall not credit as the winning pitcher a relief pitcher who is ineffective in a brief appearance, when at least one succeeding relief pitcher pitches effectively in helping his team maintain its lead.

If you don’t want to read all that, that’s how a pitcher’s Wins is tallied.  Granted, you can explain verbally in one sentence what will cover 99% of the cases.  But, MLB does go out of its way to give the complete lawyer explanation to cover 100% of the cases.  As Poz points out, this one is easy compared to the rule in determining Earned Runs.

As Poz offers:

The latter numbers have the advantage of history, which make the seem more transparent and easy to understand, though this isn’t necessarily true.

And he’s exactly right.  Those stats have the advantage of being MLB-approved.  So, while they are not the best stats, they are also not the worst stats.  If you were to grade all the MLB-approved stats, they’d come in at a B+.  They have a few silly ones as a C or D, a few would be in F.  And a whole lot that would be an A or A+.

The problem is that the internet is filled with people coming up with stats, and the internet, as a whole, gets say a C or D.  If you rely solely on Fangraphs or B-R.com, you’ll get your A- or A.  People don’t want to spend time figuring out who they should trust, and so, MLB-approved stats carries the day.

Anyway, Poz also says this:

Well, fortunately WAR is now easily available on the awesome Fangraphs Web site and also at my lifeblood site, Baseball Reference. Unfortunately, it seems like WAR is figured somewhat differently at the two places, which really hurts the brand. I wish the smartest baseball people could get together and compromise on stats like this.

Is this true?  I’ve come to calling Fangraphs version as fWAR, and Rally’s (now B-R.com) as rWAR.  That’s just for me, since I have my WAR (ostensibly would be tWAR).  The key is they all follow the identical framework, and differ on various components.  This to me is a plus, not a minus.  Would it help if Fangraphs were to call it Win Value, and Rally were to call it Wins Over Replacement, and I were to call it Wins Over Willie (W/W, in honor of Willie Ballgame)? Not to mention BPro has WARP, but which utilizes a somewhat different framework, though it may be closer to my version.

I think there’s order in the marketplace with the WAR term being used to establish the framework, and then the little letter in front to establish the implementation version.  That’s me though.  I may be in the minority (I usually am in most things).