Sunday, December 08, 2013
Lists, lists, lists
These results require much greater context.
First, the understanding that RE24 likely has a correlation of .99 if you simply did RA9 - lgRA9. RA9 is runs allowed per 9IP. You could do ERA - lgERA as well.
So, once you realize that RE24 is simply RA9 relative to league average, as a differential, the problems become obvious. A run in 1998 is alot easier to come by than in 1968. What you should do is DIVIDE RA9 by the lgRA9. Which, basically, is ERA+.
And if you did ERA+, we see that Pedro is #1. But, that by itself doesn't allow us to make any CONCLUSIONS as to who is the best. That list was simply a list, and it should be treated as a list.
Many of the comments here seem to impugn RE24, when the fact is that RE24 is simply a tool that was not properly used.