Thread: Mlb hof tracker
View Single Post
  #21  
Old 12-24-2016, 12:54 PM
nat's Avatar
nat nat is offline
Member
 
Join Date: Jan 2016
Posts: 925
Default

If you think that WAR is inaccurate, what you need to do is demonstrate that it's inaccurate. What part of the model is wrong, and why? And then maybe help us fix it.

In a nutshell, it takes discrete events (singles, strikeouts, etc), looks at how each of these events effected a team's run scoring/preventing over some specified stretch of time (whatever stretch of time you're interested in), and then converts a player's discrete events into expected runs. Which are then converted to expected wins (given how many runs you needed to produce/prevent in the period under discussion), and subtracts the number of wins a AAA player would have contributed. What's wrong with that?

Or if there's not a philosophical problem with it, perhaps there's a problem with how it's implemented? We've got some really smart people working on it, but checking their work never hurts. You can look up the equations and go through them yourself.*

*For the record, I think that there is a problem of this sort. I think that WAR systematically over-rates relief pitchers, because it includes "leverage" into its calculation for pitchers. Basically, it says that preventing a run in the ninth inning is more important than preventing one in the first. Maybe there are other such problems, if so, let's find them.
Reply With Quote