THE BOOK--Playing The Percentages In Baseball

While I admire Josh’s work, honestly, I had never heard of the proposition that relievers get “cold” in the dugout, and I thought the results (that, as a reliever, if you throw a few pitches in one inning, your stuff is basically the same the next inning) were obvious.

The reason that managers go to another reliever even after one reliever only throws a few pitchers has nothing to do with any belief that sitting on the bench affects a reliever’s pitches.  It is either to bring in a pitcher of another hand, because the new pitcher is better (and the situation is now higher leverage), or because of “set roles” (i.e., the set-up man automatically comes in in the 8th and the closer the 9th).

I don’t know about other teams so much, but that was a common excuse for the Joe Torre-era Yankees. He didn’t like bringing Rivera in for the 8th inning and having him sit on the bench getting cold. I don’t have any source for that other than old post-game interviews and such.

If that is a concern with some pitchers and some managers, of course we have the old selection bias problem here, in that the only pitchers who do pitch multiple innings, by and large, are the ones whom the managers deem capable of doing so.  We don’t get to see much data from the ones whom managers think get “cold” while sitting on the bench (and thus rarely use them for multiple innings).

I won’t parade myself as any kind of expert on physiology, but “getting cold” and recovering from fatiguing work are two different processes.
I don’t think Josh distinguished them.

#4, I agree.  I am not sure what he is actually looking at other than whether the second inning after a short first inning is any different than the first inning.  As I said, there was no reason to think that it would be.  I guess it is nice to see that it isn’t.

It is also nice to see that after a long first inning, velocity is down the next inning.  That makes sense as we know that pitchers throw a lot harder when they are only going to pitch an inning or so.  If 25 pitches is not enough to reduce velocity, then starters should probably throw harder than they do, relative to when they relieve.

I don’t think that any conventional wisdoms were debunked or strategies proven to be misguided with this data.  As I said, I don’t know of many managers who think that a reliever is not going to be effective after throwing less than 15 pitches in a first inning.  There are other reasons why certain managers do not like certain pitchers to go more than one inning, and the idea that they get “cold” on the bench (and are therefore not as effective in the next inning) is rarely one of them, IMO and in my experience.

Joe,

Yes you are right there are some potential bias issues but I don’t know how to go about removing them in the data.

MGL,

You are also right that no conventional wisdoms were debunked this article check back in a couple weeks as hopefully some nice debunking will take place.

josh,

if timestamps are now appearing in the XML pitch records, you could identify long and short inning breaks pretty exactly. I would guess muscles cooling down would vary linearly on time [anybody know differently?]. You’d want to control for fatigue and adjust or control for ambient temperature too; with a static(or maximum) 8(?) warmups when returning to the mound, there may be a delay after which a pitcher can’t fully “re-warm” before throwing his next pitch. The fur ther from fully warm, the more pronounced the effect.

sorry, lack of coffee sometimes makes me stupider. Muscle cooling would not be linear, it should follow laws of thermodynamics, which i’m sure you understand much better than me ...

The time stamps are kind of hit and miss and I haven’t integrated them in to my code but you are absolutely right.  It is very possible that a long inning of sitting on the bench has an effect on pitchers.