Baseball stats for dummies part 1: Win Shares

[JET15]

Win Shares seems to be the Bill James attempt to create the super stat to end all stats, but it really hasn't seem to have caught on.

 

The Hard Ball Times has the most recent data on it and in a very nice format, but it must be too early to put 2006 win shares out there: Hard Ball Times

 

They also link to some good explanations on what all goes into the stat. It's really too complex for mere mortals like me, but I like the concept:

 

"Bill James invented Win Shares as a simple way to compare baseball players. The idea was to develop a statistic that allows you to compare shortstops and outfielders, starters and relievers, relievers and shortstops, etc. Since the win is the ultimate measure of success, James developed a stat that measures each player's contribution to his team's wins, or Win Shares."

 

So why haven't Win Shares caught on? Could it still? Will it ever replace the old school (averages/totals) or even some of the new school (OPS/OBP)?

[rluzinski]

James made alot of seamingly arbitrary decisions when constructing the methodology for win shares. Some have also questioned much of it's logic. How James dishes out defensive win shares is especially troublesome. Still, it's interesting enough to look at. You can find an insanely detailed description of the Win Shares methodology HERE.

[JET15]

Actually, the idea of having one fielding rating is really nice because it's so hard to do. Now, I say it's nice because I haven't read John Dewan's Fielding Bible or any of that stuff. Quite honestly, I'm interested in Win Shares because I'm so lazy. A stat that crosses era's and positions and stadiums... that would be sweet if it was reliable.

 

Sounds like you're not sold, rluzinski. You seem like one of the most knowledgeable people on this forum.

[rluzinski]

The best defensive metrics use play-by-play data that shows were the ball was actually it and how hard. PMR, Dewan's +/- system (used in the fielding bible) and UZR come to mind. James' Win Shares just uses put outs, assists, etc.. for his defensive win share calculations. That kind of approach is very faulty because it can't even measure how many opportunities a fielder had.

 

How James calculates the value of defensive wins shares relative to pitching win shares, is very arbitrary. He just assumes a constant ratio IIRC. James has done so much outstanding work in the field of sabermetrics but Win Shares doesn't seem to be one of them.

 

Here's some more information, straight from the horse's mouth, when James introduced Win Shares at the 2001 SABR convention in Milwaukee:

 

LINK

[JET15]

Cool. I suppose you were there. You are the man

[rluzinski]

Cool. I suppose you were there. You are the man

 

I didn't even study baseball statistics then. Brewerfan .net turned me onto it around 2003.