Add new comment

1) The linear assumption is wrong
People think offense is cumulative and linear. Add 8 guys worth 20 RAA together than the team will score 160 RAA on average if they play enough seasons.
I believe this is wrong in the case of extreme players...in fact I even ACCOUNTED for why this would be wrong when I invented the supermargin concept in my own value methods. Players near the middle of a talent bell curve are roughly additive only because they're on the part of the cuerve that looks very much like a line. Extremely bad hitters take away opportunities from other hitters and screw up the linear weight value of the events for those other hitters even when they get opportunities. See...every single is worth 0.45 runs to a sabermigo doing the WAR calculation, but if you have three Endy Chavezes on your team, that team's average single is going to be worth more like 0.30 runs. Do a monte-carlo simulation and prove it to yourself if you have to.
2) The position adjustments are wrong
I don't do positional adjustments because I think they're based on pixie dust and fairy magic. IOW, I think where they get the baseline for each position is random and arbitrary and based only on where teams happen to play the 14 guys in a 14-team league that get the bulk of the PA. As Doc has pointed out...first base, for example, had 10 guys that are brutally awesome and 10 guys that are mediocre and they're fighting for 14 jobs. The burtally awesome guys bias the average and thus the replacement level, but they don't represent the expectation for the team with the mediocre player. I think positional adjustments are a bad idea and must be scrapped.
3) Our understanding of how to combine offensive and defensive value is wrong
Tom Tango led a movement to get people to think that there's no such thing as a replacement level fielder...or a replacement level hitter...only a replacement level total player. So he argued that RAA for offense and defense should be added together first and then a replacement level be found. That isn't the right approach IMHO. I think value is value FIRST...I think you have to figure out where the replacement levels are for each skill and then add the results...or you get whacky things like Endy Chavez being worth 25 runs somehow because his defense is way above average and his offense is replacement level and when you combine them, you get a weirdly high number.
All three of these factors create a perfect storm of bad ideas that cause our evaluations to be way...way wrong especially for guys like Mike Morse.