Apparently there is a great deal of confusion and consternation over Robin Ventura having Alejandro De Aza sacrifice bunt with runners on first and second and no one out in the 7th inning of yesterday's game. This is what happens when dilettantes in the press try to stir up the sabermetric debate. To wit:
But Ventura smiled when asked about how sabermetricians might frown on his bunt strategy and giving up an out.
"Well, they're not sitting in my seat, either," Ventura said. "There's a lot of those guys out there but they're not sitting in this seat and they're not sitting in this dugout. So it's a different feel when you're a player or a manager than it is just to sit there and right numbers down on a piece of paper."
And from Mr. Roosevelt's summary of some sports talk radio:
Boers and Bernstein were prattling on about how you don’t give away an out (with regard to the de Aza bunt) and were laughing about how much Hawk loved the bunt. One asked the other, "Why did he love it so much?" and the other responded, "Because he doesn’t understand win expectancy," so I looked up the win expectancy on FanGraphs both before and after the bunt and it was identical. I’m sure part of that had to do with having a three-run lead in the late innings and therefore each individual play would have relatively little impact, and to be fair the radio guys may have meant run expectancy rather than win expectancy but if you want to sound smart complaining about that bunt in that situation, using win expectancy as an argument is not the way to do it.
Since I'm apparently an authority on this topic due to some past writings, let's clear this up.
Mr. Roosevelt is right that run expectancy, not win expectancy, is the relevant inquiry here because win expectancy isn't going to move much in this game situation. (Using reasonable assumptions, the successful sacrifice increased the White Sox' chance of winning from 0.9585 to 0.9593.)
Late in a game and with a lead - particularly a lead of three runs - it's okay to play for one run; meaning, increase your chances of scoring a run as opposed to increasing your chances of scoring many runs with the commensurate risk that you won't score any runs. That's because the value of two or more additional runs isn't really that much higher than one additional run because the other team is already behind.
As you can see from this run expectancy chart (the chart in the middle of the page), the chance that a run will score at some point in the inning is increased slightly with runners on second and third and one out from the chance with runners on first and second and no one out.
Yes, as the chart at the top of the page will tell you, this reduces the expected average number of runs scored in the inning. If he was swinging away, De Aza might have not made an out and done all sorts of other wonderful things like take a walk or hit a bases clearing double. Or he might have still made an out but done all sorts of other less wonderful things like failing to advance the runners or hit into a double play.
What Ventura did was a low reward but also low risk move. It didn't work out due to some poor baserunning. But that doesn't mean it wasn't an appropriate tactical choice to make given the game situation.