Jimmy is hitting for a high average, but he still has some work to do (look at the walks and strikeouts). He’s a big league prospect and his time will come.

Tango at The Book blog does a pretty good take down on a blog piece attacking the sabermetric critique of sac bunting. The target is an article called "inconvenient truth about sacrifice bunts," which throws down the gauntlet with "Stop the presses....that 'everything you know is wrong' twaddle that the neo-sabe movement has been overselling is ready to take one on the chin." A broad claim like that makes you think that we will be enlightened to find that giving away outs is a winning ticket.

I'm not writing this as an evaluation of the pros and cons of the sacrifice strategy. Sabermetrics is rightfully skeptical of the overuse of sacrifice bunting. For many years, sacrifice outs were handed out like candy in situations when their use would reduce the team's run expectancy. However, modern sabermetrics recognizes that there are some situations when sacrifice bunts are useful, and may well be justified occasionally in a game theory sense. So, I don't really agree that sabermetrics is as rigidly anti-bunt as the "inconvenient truth" author says.

My interest here is more about Tango's effectiveness in identifying the logical fallacies in the article.

The article says that sabermetrics fails to look at the winning percentage of games in which the sac bunt is employed. The author says "what is so shocking" is that the win-loss record is so good, "more than good, in fact." The winning percent for teams that employ the bunt one or more times in a game is 64%, which evidently proves that it is a winning strategy. The conclusion is that the sac bunt is associated with an elevated winning percent. The author goes on to argue (rather unclearly, in my view) that sacrifice bunts are neutral game elements.

Tango should have begun his response with Samuel L. Jackson's "Well allow me to retort."

First, as the article itself notes, sac bunts are more often used by a team that is already winning than one that is on the losing end of the current score. (62% of sac bunts in non-tie situations are undertaken when a team has the lead.) If the bunting team already had a lead in the game, is it any surprise that the team shows a winning record, given this advantage? As Tango says, "sac bunts happen to occur when the team already has a good chance of winning."

Next, Tango addresses the article's claim that more runs are scored when teams use the sac bunt:

That is, in the PA of the sac bunt, and all PA that follow, the average team scored 1.031 runs.

How is that possible? Well, if we look solely at the base-out situation presented, and regardless of whether the batter bunts or not, we expected 1.058 runs to score. So, this is why we get a lot of runs scored in innings when you have a sac bunt: it’s because you happen to have a runner or two already on base! And in Table 11, we get to the blogger’s point. We see that following a sac bunt, the team wins 62.8% of the time. But before the sac bunt, based on the inning, score, base-out, the chance of winning was 63.8%. (emphasis added)

Tango goes on to point out several other plays where the same fallacy of "association with winning" exists. For example, a team wins 56% of its games after a caught stealing. But this doesn't mean that getting caught stealing is a good thing. It just reflects the fact that teams which have a lead are more likely to attempt a steal. Receiving an intentional walk is the event with the highest subsequent winning percentage. Again, one of the main reasons is that a batting team receives an intentional walk most frequently when it is already heavily favored to win.

Oh, by the way, Astros' manager DeFrancesco said that Scott Moore's sac bunt in the first inning of Sunday's game was Moore's idea, according to a Zachary Levine tweet. It's almost like Tony D doesn't want a sac bunt to be his first introduction to Astros' fans.