THE BOOK--Playing The Percentages In Baseball

Forgoing clutchiness for a moment, I just wanted to see if this looks any more real when I type it out:

Bonds, 2004: +12.7 wins.

Nope.  Still doesn’t seem real. 

*insert “well it probably wasn’t” comments here*

Would love to see a leaderboard on this site somewhere, broken down by season, by SP, RP, and batters.  And perhaps another by fielding position.

Other notable numbers: Gagne’s 2003 WPA of 731.  Besides Barry’s stats that were influenced by the ridiculous IBB’s, this is what I’d think is the most “untouchable” WPA total for a season.

Barry’s stats were NOT influenced by the IBB.  If the IBB had very little impact to the win probability when they were issued, then Barry will achieve very little benefit.

I would even say that Bonds’ WPA per PA, excluding IBB in the numerator and denominator, would likely match his overall WPA per PA.

I’m confused for a couple reasons.  First, I’m guessing that an IBB can never result in a loss of WPA.  And second, I’m guessing that many of Barry’s IBBs occurred in high leverage situations.  I guess you’re arguing that the slow accumulation of positive WPA is matched by the moderate losses and great gains in WPA when he makes an out and homers, respectively, in these situations?  I guess I just intuitively think the opposite.

I could have sworn someone created a table a couple years ago for determining when to walk Barry and when not to.  (I think it was you tangotiger, but I could be wrong).  And if memory serves, it seemed to me that Barry got walked more than he should have.  Which would indicate to me that, had he been pitched to in these situations, the pitching team would have had a better chance to win the game and thus Barry’s WPA would have gone down.

Yup, that was me.

Let’s take Barry’s WPA from 2004: is +12.7 wins on 610 or so PA, including IBB, for an average of +.021 wins per PA.

In 2003, it was +.014 wins per PA
In 2002, it was +.016 wins per PA.

So, Barry is +.017 wins per PA, more or less.

How much is an IBB worth, in terms of wins per PA?  The typical IBB from 1999-2002, for all players, is +.010 wins per PA (see table 11 of The Book).  However, Barry receives more atypical IBB.  His IBB actually moves runners over, or occurs in other situations where it would seem rather strange.  Fangraphs will soon start publishing the WPA on a play-by-play basis.  I would not be surprised that if we add up the WPA on Barry Bonds’ IBB, that the win value of such IBB will be around +.017 wins per PA.

That is, when Barry is walked intentionally, he averages +.017 wins per PA.  When Barry is NOT walked intentionally, he averages +.017 wins per PA.

In effect, he gets IBB walked at the right times.

In reality, it gets better.  In the situations that he does get IBB, the LI is likely more than 1.0.  It’s probably 1.5, if not even higher.  So, Barry, overall, when his LI is 1.0, is +.017 wins per PA.  But, when the LI is 1.5, he’s worth +.025 wins per PA.  If issuing the IBB to Barry means a gain of +.017 per PA, then issuing such a walk, because it’s Barry, actually REDUCES Barry’s effectiveness.

Because of Barry, this is a fascinating subject.

Fangraphs.com is improving on a daily basis.  Check this out:

http://www.fangraphs.com/statsd.aspx?playerid=1109&position=OF&season=2004

The WPA is now part of the daily logs.  And sortable.  So, we see that Barry Bonds, on Aug 21 against the Mets, had a WPA of +.87 wins.  In a game the Giants lost!

I remember that one of the arguments against clutch hitters (that I used myself) was that the perception of who hits well in the clutch doesn’t match the actual data.  Here, however, we see that the guy perceived as the god of clutchiness actually has been the god of clutchiness over the last two years.  And, A-Rod, viewed as a choke artist this year, has indeed been a WPA loser.

http://www.baseballprospectus.com/article.php?articleid=5420

Joe Sheehan said this:

Opportunity plays an enormous role in determining who gets called clutch, and Ortiz does seem to have an inordinate amount of opportunities, in no small part because of the guy hitting behind him.

I’m surprised by Joe here.  Ortiz’ LI is 0.99, while Manny is 1.03.  The average Redsox hitters are at 0.99.  Joe must have the perception that because Ortiz has come through so often is because he was so successful.  Or, he’s associating clutch to RBI.

Manny and Ortiz match their SLG, but Manny has a 30 point advantage in OBP.  And even with that huge headstart, Ortiz’ WPA is +2.3 wins higher than Manny.  That is simply an enormous difference… towards Ortiz.

Thanks to the incomparable David Appelman, we see that Ortiz has +1.7 wins of Clutchiness this year, to go with this +3.7 wins of Clutchiness last year.  +5.4 wins of Clutchiness is likely the Clutchiest Clutch performance of all-time.

This year, his bat has added +13.8 win advancement (WA), and added 8.2 loss advancement (LA).  Manny is at +11.1 and 7.8.

Both these hitters have come to bat about the same number of times, and the leverage they faced has been around the same, but Ortiz has clearly been making the most of it. 

The argument for Ortiz is based on his clutchiness, and Joe pretty much sidesteps the whole issue in one sentence.  I imagine that his email box will be littered with Redsox fans.

For 2005 MVP rather than have the same argument in 2 places…
http://www.fangraphs.com/forums/showthread.php?t=53

I’d also like to note something odd here that I’ve found while looking through the “clutchness” numbers.  It is pertinent this thread but not the other.  Geoff Jenkins has been crazy this year.  He has the highest amount of clutch wins I’ve found.

http://www.fangraphs.com/statss.aspx?playerid=690&position=OF

Yup, that’s true, as fans of http://clutchiness.blogspot.com are well-aware of.

While David Ortiz continues to pile up the Clutchiness wins, Manny Ramirez continues to be one of the biggest choke artists in baseball!

http://www.fangraphs.com/statss.aspx?playerid=210&position=OF

His 5-year Chokiness score is 5.6 wins.  Ortiz’ 5-year Clutchiness score is 5.3 wins.