Thursday, June 2, 2011
A Brief Foray into Run Differentials and Component Stats
At this point in the season, the Yankees have the second best winning percentage in the AL.
Over a full season at the actual winning percentage the Yankees would be about a 93 win team, with Cleveland at 101.
Now of course over 1/3 of a season actual winning percentage can be misleading since teams might be winning or losing more games than their actual performance merits. So you can look at something like Pythagenpat to get a better sense of how well a team has played so far and what it might mean going forward.
If you do that, the Yankees look more like a 102 win team, which would be eight wins better than Cleveland and 11 wins better than Boston.
Although Pythagenpat does a pretty good job of estimating a team’s level of play, their actual runs scored and runs allowed might be skewed due to better or worse than expected performances in high leverage situations that are not necessarily repeatable going forward.
You can use linear weights batting runs to account for that. What’s nice about doing that is you can put offense and pitching/defense on the same scale if you use all the same components.
This method also shows that the Yankees have played better than any other team in the AL, and would have them at around 97 wins, five wins ahead of Cleveland and six games ahead of Boston.
We do need to be cognizant that how a team has played so far only tells us so much about how good they are now and how good they’ll be going forward. Regression towards the mean, injuries/roster changes and a whole host of other things are going to have an impact on how a team does moving forward.
But at least as of right now, the Yankees have probably been the best team in the league and the difference isn’t trivial.
All this is moot when Rafael Soriano returns to blow games though.
Page 1 of 1 pages: