from Amazon
Monday, July 21, 2008
Late-inning fielding replacement
I know I wrote a post on how to figure out if a late-inning substitution, trading offense for defense, makes sense. From what I remember figuring out, if you are trading one extreme for another extreme, then you should do it in the 9th, maybe in the 8th, and definitely not in the 7th or earlier.
If someone can point me to what I wrote, I’d appreciate it. I don’t want to bother redoing everything I did.
UPDATE: I’ll just rewrite what I think below:
The difference between a great fielder and a bad fielder is roughly .27 runs per game, or .03 runs per inning.
The difference between a great hitter and a bad hitter (at the same position) is roughly 50% higher than that, or say .43 runs per game, or .10 runs per PA.
If you replace someone for one inning for defensive purposes, you do it. Of course, we don’t know that it will be for only one inning, since extra innings are always a possibility.
If you replace someone for two innings, then it starts to get dicey. If you are replacing the game who made the last out in the 7th, then there’s a good chance that he won’t come up in the 9th. I’ll guess the odds as around 50/50. And if he does come up, chances are that means you scored more runs on the board. So, for two innings, you get +.06 potential runs with the glove against a 50% chance of losing .10 runs at a likely lower level leverage.
For 3 innings, it starts to get dicey. You get +.09 runs with the glove, and you lose .10 runs with the bat.
So, does it make sense to replace a guy in the 8th, or even the 7th inning? Well, it depends. It depends on exactly who the players are, when is the replacement due up, and what the leverage is, and expected to be.
Even if the manager limits it to the 9th inning only, and he does it with players of extreme and opposite ability in off/def, and he does it 30-40 times a year, that’s a total gain of +1 run.
For such an extreme tradeoff, to get one run in this deal, I have to question the idea of even having a “defensive specialist” on hand for such situations, when the “defensive specialist” I would like would be another guy in the bullpen so that I can get my below-average starters out of there after 4 or 5 innings.
I am gonna guess that if it is correct, that it makes very little difference one way or another. One reason that is probably true is that we (including managers) probably don’t take into consideration any offensive and/or defensive penalty (like for a pinch hitter or a DH) from coming into a game cold in the 8th or 9th innings. You are generally bringing in a much worse hitter. What if you add in a “pinch hitting” penalty? Now he is a horrendous batter (obviously sometimes and ideally the defensive replacement does not bat)
And what if there is a defensive penalty to coming into a game cold? I would think there has to be. That would be something to look at in the UZR data!