[curious]

Hi,
I have been looking for statistics in the various sports that have good predictive power. For basketball it is points per possession, for NFL it is starting field possession...etc.

For baseball I believe it is runs per plate appearance. You won't find this in google, I am developing it. A correlated stat is of course plate appearances per game. Similar to pace in basketball.

I think the data shows that runs per plate appearance on offense and runs allowed per plate appearance on defense correlates closely with games won/lost.

You have to also look at how many plate appearances the opposing pitching staff and defense allow. A pitcher gives up a plate appearance for a walk or a HBP and for a hit ball which is not fieldable. A defense gives up a plate appearance for a ball hit into a fielder's zone which was not an out or an assist.

I think that for fielding errors don't mean anything. Outs per batted ball are what is important. Some people call this the zone factor. This can get complicated because not every batted ball is fieldable even though it is in a given player's zone. A line drive that bounces off the left field wall is the pitcher's problem not the left fielder's, because even though it was hit into his zone he couldn't have caught it.

Zone factor, or what I like to call outs per batted ball are much more important than errors. For example, when Brooks Robinson was in his prime, any ball hit between the 3rd base line and the shortstop was pretty much a guaranteed out. There are other third basemen, Ryan Braun for example, who misses 30% of the balls hit into his zone. OUCH. He doesn't make a lot of errors, but that is because he doesn't get to the ball very often.

So, a team with a high runs per plate appearance avg and a high outs per batted ball average will win a lot of games.

Anyway I am turning this into a predictive model similar to the one I am using for NCAAB.

[diogee]

Thanks for sharing...looking forward to reading about it during baseball season.