I thought I might be able to whip up a quick but interesting and possibly useful metric to gauge the number of wins lost or saved due to injury from the table Jeff whipped up. I like the percent of salary measure, but I thought the total days tally might be more telling of training staff aptitude. If you assume that the team loses a 1.5 WAR/150G player and has to call up a replacement level player and .85*Days Missed = Games Missed, then we can use the table and say
.85*DM*1.5/150 = .01*GM = WAR lost to injury
I'm guessing most actual team coefficients are somewhere between .005 and .013. 1/100 is nice and round. .0088 is 33 WAR per 25 players per season. If the best players get by far the most playing time, I assume they will get injured the most. So, .01 it is. Presumably there's a correlation between cost and talent, so if a team tends to lose expensive players to injury, this metric will understate the actual wins lost. But I assume that injuries are distributed such that good expensive players get injured relatively little.
|Team||Total Days||Avg Days/Season||WAR Lost/162G||WARLAA/162G|
WARLAA = WAR Lost Above Average, so negative is better. Anyway, the White Sox are apparently good at this health stuff. I assume this explains some of the disparity between their projections and performance. It looks like team health is worth +/-3 wins or so per season, which they and the Angels made the most of. That's obviously a very significant number of wins, the equivalent of adding another above average player to the team. It would also appear that percent salary lost and this little metric correlate okay, but that's only employing the eyeball test. Thoughts?