Fun with Win Shares
You all know what Win Shares are? Bill James invented them a few years ago as a way to express the total production of a player, in terms of the amount of wins his contributions have produced, in and of themselves. I don't really know how he came up with it, but anything that James does is highly significant, as he is one of the top minds (if not THE top mind) in the world of baseball statistics, as well as the game of baseball itself.
A Win Share signifies that that player contributed to one third of a win; therefore, if a player has 9 Win Shares, it means his total production has meant 3 wins for a team. Now I wouldn't exactly consider this statistic gospel, as it's just an approximation, but it's as relevant as any other statistic that tries to translate other statistics into a summary of a player's true value, such as VORP or Runs Created or whatever.
Anyway, Win Shares also has it's own method for defining a player's defensive value: I guess every contribution on the field is given value. So it's more than just an offensive/pitching statistic. This gives an edge to players who actually play the field, and hurts Frank Thomas and David Ortiz and all other DHs. It also benefits players who play up-the-middle positions (catcher, SS, 2B, CF) as they are involved in more plays, and have more value for doing well too.
This was actually a poor description of the statistic; if you really want to understand Win Shares, James wrote a whole book about it.
ANYHOO: over at The Hardball Times (http://www.hardballtimes.com/thtstats/main/index.php?view=winshares&linesToDisplay=65&orderB y=total&direction=DESC&season_filter%5B%5D=2006&pos_filter%5B%5D=All&Submit=Submit),
they have Win Shares updated through August 29. It's great, all the stats are sortable, so we can see who's the best catcher in the NL, or who's the best defensive shortstop, or who's the best hitting pitcher, etc. Some notes while searching through the stats:
- Best player in baseball this year (by total Win Shares): Carlos Beltran, with 34. He actually has the third most Batting WS, but his defensive value is so great he has 4 more WS than anyone else; he's rated as the best defensive outfielder, by far. Players with high defensive value are given their dues by Win Shares.
- Following Beltran are Albert Pujols (30 WS), Miguel Cabrera (also 30 WS), Alfonso Soriano (28 WS) and Lance Berkman (27 WS) - making the top 5 most valuable players in baseball this year National Leaguers. Coming in at #6 is Joe Mauer; again, his defensive value puts him over the top of other ALers (although Manny Ramirez and Derek Jeter are right behind).
- No A's are in the top 50. Coming in at #56 is our top player so far in 2006: Jason Kendall. That's right, our top player this year (according to this metric) is Jason Kendall. Once again, our top player this year is Jason Kendall. Jason Kendall! He has 18 Win Shares, and in actuality, the difference between Kendall's contributions this year and those of Zito and Swisher (both at 16 WS) is pretty insignificant, but it's still there. He rates so high mainly because...
- Kendall is the top rated defensive catcher so far this year. Once again, KENDALL IS THE TOP RATED DEFENSIVE CATCHER SO FAR THIS YEAR. Got that? In case you were wondering, he was nowhere near the top in Defensive Win Shares last year. I believe I read somewhere that defensive Win Shares for catchers depend highly on the amount (and maybe percentage) of baserunners thrown out, so that explains some of it; more of it is explained by the fact that, I'm sure, he's spent more inning s behind the plate than any other catcher, and therefore has had more opportunities for Win Shares. Plus, he's barely ahead of Pudge Rodriguez. Still, they calculate that Kendall is the 4th most valuable catcher in baseball this year. Feeling better about the $10 mil he's getting this year?
- On another defensive note, they rate Mark Ellis as the top defensive 2B in the game - awesome! We all knew he was great.
- Chavez is (only) the 3rd best defensive third baseman, behind Joe Crede and Mike Lowell. Rounding out the infield, it's unfair to compare Crosby and Swisher to others at their positions, because they haven't played full years there (plus they list Swisher as an outfielder). Our outfield defenders are not thought of highly here. Center fielders tend to get the most defensive WS, but Kotsay is way down on the list, and Payton and Bradley are further down.
- Our top hitter: Frank Thomas, at #46 (15.5 Win Shares). Yeah, our offense has sucked this year. Swisher is next at #67 (13.4 WS).
- Now the other facet of of secret to winning this year: pitching! The top 5 in pitching Win Shares are Johan Santana (20.8 WS), Roy Halliday, Brandon Webb, Chris Carpenter and Carlos Zambrano, but Zito (16.4 WS) and Haren (15.6 WS) rank 11th and 13th. Street and Blanton are the A's next best, coming in at #s 40 and 42.
- Anyway, it's a really fun stat and I recommend checking it out. This stat really reflects how we all seem to feel about the 2006 A's, that it's waaaaay more about the pitching and defense this year than the offense (although I guess that's kind of obvious).
0 recs |
4 comments
Comments
win shares were chatted about
This is causing Win Shares to go in my mind from Complicated Hacked Together Stat I Don't Understand That Seems To Correlate Sort Of To Wins to Highly Questionably Crap.
I mean WTF. According to this, Kendall is the most valuable fielder in ALL OF BASEBALL, leading Pudge by a slim 2% and the 3rd best player, Mauer, by a massive 26%.
(You can see by the league leader list that Catchers get a lot of fielding win shares... )
The thing I don't like about Win Shares is its complete hackish nature. It looks like it was put together by a government committee (and in fact, it resembles the formulas used to decide on public school funding distributions)... lots of ad hoc, arbitrary tweaks to try to make it fit data better. Most other saberstats have some meaning when you think about them for a while like SLG, OPS, WHIP, win expectancy added, or ERA+, etc. Not Win Shares. Without that transparency, unless there is some overwhelming statistical predictive power, it's not a very useful formula.
For instance, James's Pythagorean Formula is funny looking and the exponent seems arbitrary (mathy folks should see Sal and andeux's posts on Weibull distributions though), but it's damned predictive, so you have to acknowledge its power. I would love to see a simple study comparing Win Shares for predictive value compared to other stats, including unvarnished simple winexp (which should not be particularly predictive).
by Apricot on Aug 31, 2006 10:00 PM PDT reply actions 0 recs
FIRE MACHA NOW!
-Nick Swisher
by kaweahkaweah on Aug 31, 2006 11:35 PM PDT reply actions 0 recs
Jermaine Dye's fielding?
Does Dye have a general rep as a poor defensive outfielder? I don't recall that as a big issue when he was here. (But then, when he was here our expectations on defense weren't high.)
by matthias on Sep 1, 2006 12:22 PM PDT reply actions 0 recs
weird
Note that Win Shares is methodologically weakest when it comes to fielding. You can see a brief version of Win Shares computations at http://www.baseballgraphs.com/main/index.php/site/details/
First and foremost, Win Shares does not use play-by-play data. For outfielders, one uses
Outfielders: Putouts, Team DER, Arm Elements and Assists and Errors
which others have critiqued.
Honestly, I think most public defensive metrics are quite inaccurate, to the point where it's almost not worth doing it unless you do play-by-play with careful charts (I think UZR does?), and any really careful studies like that are proprietary and being paid big bucks but the actual teams...
by Apricot on Sep 1, 2006 4:31 PM PDT up reply actions 0 recs

by 




















