from Amazon
Tuesday, April 22, 2008
Fastball speed and movement, while resting
Ah, just what I’ve been thinking. In The Book, going through every single relief appearance in four years, I was unable to find any impact to a reliever being overworked (p. 231-234 I think). That’s not to say that the effect isn’t there. It’s just tough to find in “output” stats, since random variation causes alot of pollution. But, “input” stats, like fastball speed? Far less random variation to be sure. And Josh shows the effect exists on the third day of a 3-straight day effect. But, is it true? I don’t know. He doesn’t show if it’s the same pitchers in each bucket, and therefore, might simply be a selective sampling issue. You need to have a baseline to compare against. And, he also looks at movement (sinking action).
Good job…
Sorry to repeat what Guy and Tango said, but you CANNOT do this kind of analysis without controlling for the pool of pitchers in each bucket. As they both point out, the pool of pitchers in each bucket is probably very different, so you CANNOT say that pitchers lose or gain velocity or movement after x days of rest or pitching on y consecutive days, regardless of what the data shows.
The best way to do this, as we do with aging studies, is to simply keep track of plus or minus velocity (the “delta method") as compared to some baseline. It does not matter what baseline you use, as long as you tell us what it is, and you use the same baseline for each bucket.