Should Nick Green have bunted last night with nobody out and runners on first and second in the bottom of the 11th inning?
My initial reaction to these sorts of questions is always no. Don't give up the out. And for the Red Sox, that's almost always the correct answer, which is why you see the Sox bunting so infrequently. But Green presents a different situation in that, for the Red Sox, he is a uniquely bad hitter. The following is reprinted from what I wrote in the game thread below.
In the 2002 run environment (the closest numbers I could find via a quick Internet search), with runners on first and second, zero outs, a team should expect to score on average 1.511 runs in that inning.
If a hitter successfully sacrifices them to second and third, the run expectancy drops to 1.358.
Of course, if he strikes out or pops out, run expectancy drops to 0.936, and if he grounds into a double play, advancing the lead runner to third with two outs, it plummets to 0.363.
The question is how much confidence do you have that your hitter will avoid the double play and at least advance the runners. Nick Green strikes out 21 percent of the time and walks six percent of the time, so the chances are good (73 percent) he will put the ball in play. Of those, 40 percent are ground balls and 60 percent are fly balls.
Meaning that in any given situation, Nick Green is likely:
To hit a fly ball 44% of the time.
To hit a ground ball 29% of the time.
To strike out 21% of the time.
To walk 6% of the time.
Six percent of Green's fly balls have been home runs, translating to less than 3 percent of his total appearances. Of Green's remaining balls in play, 30 percent have been falling for hits this season, meaning Green had an overall 19 percent chance of getting a base hit (not counting platoon advantage/disadvantage, park effects and other important considerations I don't have the time to figure out right now, but one can assume, given the Sox were in Fenway, that Green's chances of getting a hit were actually higher).
Meanwhile, in GIDP situations, Green has done so at a 16 percent clip. So revising our percentages and remembering that this is context neutral for type of pitcher, ballpark, etc.:
Green had a 21% chance of striking out,
A 19% chance of getting a hit short of a home run,
A 16% chance of grounding into a double play,
A 6% chance of walking,
And a 3% chance of winning the game with a walkoff.
That's a 28 percent chance that Green's at bat increases the run expectancy for the inning, and a 37 percent chance that he decreases it worse than he would have done so by bunting. That doesn't count infield flies, which Green has hit at a nice rate of 19 percent of all his fly balls. If we assume fly balls make up most of the remaining 35 percent of plate appearances, then that means another 7 percent chance that Green worsens the run expectancy by failing to hit the ball out of the infield.
So when Green stepped to the plate, he had a 28 percent chance of increasing the club's run expectancy in some form or fashion. (His OBP is actually .311, so I might be cutting him short a little here by ignoring things like HBP, infield hits, errors, etc.) He had a 44 percent chance of drastically lowering the run expectancy by failing to at least advance the runners. And he had a 28 percent chance of advancing the runners with a fly ball out.
So bunting is actually pretty defensible in that situation, given how bad Green is with the bat. He had a much greater chance of doing even more damage than bunting, and while I agree that giving up outs is not something you should really ever do, if you're going to do it, that's the player and situation with which to do it.