Last week I wrote that upon investigating the play-by-play of the entire 2012 NL Home Season, I found that a marginally greater proportion of teams that did not sacrifice bunt in bunting situations actually ended up scoring later in that inning than teams that did. Specifically, I found that when a teams that are given a sacrifice opportunity (defined as a situation when there is at least one runner on base but the bases are not loaded, and either there's nobody out, or there's one out and the pitcher is at the plate), when they sacrificed they ended up scoring later that inning 47.4% of the time; otherwise they scored 51.8% of the time. The difference is significant at the 90% level.
Although this look is narrowed from simply examining differences in Run Expectancy tables (for reasons explained earlier), it still is highly aggregated and therefore might be missing some intricate detail. What happens, in particular, when we look at who is doing the bunting? It certainly makes sense (i.e. it has certainly been argued) that sacrificing with weaker hitters (especially pitchers) makes more sense than having other guys do it.
Well, so here's a table detailing the chances of a team scoring when using or when not using a sacrifice, depending on the batting order position of the player: guys hitting 1-8 or guys hitting in the 9-hole.
|Did Team Bunt?||#1-8 Hitter||#9 Hitter||Overall|
This is no typo. When a team sacrificed using their #1 through #8 hitter, they ended up scoring runs more often than when they didn't; and when a team sacrificed using their #9 hitter, they ended up scoring runs more often than when they didn't; but overall, when a team sacrificed, they ended up scoring runs LESS often than when they didn't. In fact, with the exception of bunting from the #8 hitter, teams that sacrificed with their n-hole hitter scored more often than teams that did not sacrifice with their n-hole hitter!
So, you may ask, how can this be? Are maths failing us? Are we looking at a mirage?
I think it makes perfectly good sense that teams that sacrificed with their #8 guys fared worse than teams that didn't, but the other stuff...?
Well, there are two things going on. First of all, when a team is in a potential sacrifice situation with their #9 hitter (most often, their pitcher) due up, they aren't really in that great a scoring situation to begin with. They scored less than 40% of those situations, which is well worse than when teams had anyone else up. This makes sense. The other thing is that teams were MUCH more likely to employ the sacrifice with their #9 hitter (again, that makes sense). Specifically, 606 of the 974 bunts executed were put down by #9 hitters (62% of them). By contrast, only 973 of the 8362 (12%) non-bunts in bunting situations were performed by #9 guys. From a different perspective, only 4.7% of #1-#8 hitters employed a sacrifice when in sacrifice situations, while 38% of #9 hitters sacrificed in sacrifice situations.
Bunting is thus most often used to make the "best out of a bad situation". Not that having runners on base with less than two out is bad, but if the pitcher is up, it's not a particularly great situation either (and having the pitcher bounce into a double play WOULD be bad!) Since that is the time that bunts are most often employed, that's why the overall results are so much worse. As an (admittedly morbid) analogy, the survival rate of people who are given chemotherapy treatment is very likely lower than the survival rate of non-chemo patients among the population as a whole, even though the rate is higher when just considering cancer patients.
What do we take away from this? At this point I think the most important thing to note is that NL managers apply the sacrifice judiciously. So while the chart above shows a 86% chance of scoring when a #3 hitter sacrifices (as opposed to a 64% chance when the sacrifice is not used), it's important that #3 hitters only sacrificed 7 times all year in 2012. And #4 hitters only sacrificed 3 times. These were no doubt special occasions, or surprises, or maybe they were REALLY just failed attempts at bunt base hits -- four of those seven sacrifices from #3 hitters came in the first inning. Under ordinary circumstances most managers would frown upon their cleanup hitters taking it upon themselves to sacrifice. But there may be times when the hitter, the matchup, the score, inning, and all the other considerations of a game situation make it make sense for NL managers, those among the "30 dumbest people in the world", to signal for a bunt; and in so doing these managers have actually improved their chances of scoring in those certain special occasions.