Add new comment

...which means that over the long haul...it should even out...but...we all know how that assumption goes. Not saying what I try to do will always be a perfect measure...but I believe CERA can be measured just as accurately as, say, a park factor can be - or a platoon split. :) It just takes a certain number of reps to get a clear picture.
You don't have to worry too much about the defense behind a pitcher if you measure everyone against his team's total defense - it won't account for the random luck of who gets who as their fielders, but those variables are likely small.
The approach would be the runs scored for a game by the opposing team are equal to:
The league average R/G + the dR/G for the opposing team's offense + the dR/G of your side's defense + dR/G of the home plate umpire + dR/G of the wind direction relative to the park + dR/G of the starting pitcher as a whole + dR/G of the starting catcher.
dR means delta Runs (as in the skew from average by the component skill)