This afternoon, I showed that the pitcher win values are actually fairly decent predictors of the next year’s win values, with an average year to year correlation of around .63 for the last four years. How does this compare to the hitter win values we publish here on the site? They’re pretty comparable. How comparable?
2004 to 2005: .59
2005 to 2006: .63
2006 to 2007: .64
2007 to 2008: .66
That’s also an average year to year correlation of .63 – the same as we found for pitchers. Just knowing a player’s prior year win value, you’ll have a rough idea of what his following year’s win value may be. Now, a the advanced projection systems (CHONE, PECOTA, etc…) will do a better job of incorporating more data and weighting it appropriately, so we’re not suggesting that single year win values replace those systems for projecting future performance. However, it’s a good sign that the win values correlate fairly well from year to year.
So, hopefully, the posts we’ve done the last few weeks have helped you understand what the win values system is telling you, and the transparency we’ve tried to apply to the system should allow you to trust the results. The win values system captures player value very well, and predicts future win values for both batters and pitchers with solid reliability.
We’re not claiming this system is perfect. This isn’t the perfect single number metric that sums up all player value with no error that people seem to want. But, it’s pretty darn good, and as good or better than anything else out there. Right now, if you want to know how much a player is worth to his team, Win Values are your best bet.