from Amazon
Friday, December 23, 2011
Sending runner on 3-2 count
This was from Tippett, and a followup.
In a blog entry the next day, we offered the opinion that this was a poor decision with the league leader in strikeouts on the mound. We didn’t get into the details then, but we will now.
In last year’s essay, we pointed out that 11.4% of all plate appearances in double-play situations result in a ground ball double play. Not all of these plate appearances see the batter put the ball in play, though, and when we subtract the walks, hit batsmen, and strikeouts, it turns out that 15.4% of balls in play produce GDPs.
If you DON’T send the runner on a 3-2 pitch, the chance of getting a ball in play that produces a GDP is a little over 5%. That’s because 15% of balls in play result in a GDP and batters put 3-2 pitches in play about a third of the time.
So what are the chances of a double play if you DO send the runner on a 3-2 pitch? Last year, we estimated that probability at about 5.5% based on the frequency of strikeouts, a typical caught stealing percentage, and the small chance that a ground ball up the middle will produce a GDP even with the runner going.
In other words, the overall averages showed a rough balance between sending and not sending the runner, leaving the decision to an evaluation of the skills of the specific players involved.
But that was for the case where the only runner was on first base. What about the first-and-second situation?
I just quickly read through Part I. I think he really screwed up the analysis. First of all, he includes foul balls in order to estimate how often the ball is put into play. Foul balls should be ignored, obviously. They have no bearing on the analysis (other than tiring out the runner).
Secondly, I think he forgets about base runner advancement on the K, since he concludes that for an average base stealer (say 65-70% SB rate) it is a BE proposition to send the runner. If the CS and the SB on the K cancel one another out, then any benefit from staying out of the DP is a bonus. He says that since the CS/K DP occurs at the same rate as the GDP (if the runner does not go), then it is a BE proposition. Again, that ignores the value of the SB on a K.
Unless I read him wrong.