There is a place for math, but in this case, the math is fuzzier then it first appears.
First, the math assumes that all base/out situations are created equally. That 2nd base, no out, with a hard throwing reliever against the bottom of the order is the same as 2nd base, no out, middle of the order against somebody's tiring 5th starter. That LH/RH don't matter for pitcher or hitter. None of this is ever true. This is a case where averages obscure the truth, rather than reveal it. There's an old saying...a statistician will look at a guy with one foot in boiling water and one frozen in ice and conclude, on average, the guy must be pretty comfortable.
Second, having seen these studies before, I believe the math shows that in certain situations, sac buntingactually increases the chances of scoring exactly one run, while lowering the chances of scoring multiple runs. And I agree with RB...there are situations when scoring one run is so important that I'd be more than willing to lower my chances of two or more to increase my chances of one.
And third, I don't even care if Kepler bunts there. A sac bunt attempt isn't guaranteed to advance the runner either. But if the team isn't going to bunt in that situation, then the hitter absolutely, positively, has to sell out to hit the ball to the right side, to maximize his chances of advancing that runner to third. In this case, Kepler swung wildly at the first pitch, a pitch that was very "pullable," and then got beat on a pitch away, which he had little chance of pulling, and popped up weakly to the left side.
That's a failure to understand the situation, IMO, not just a failure to get the job done. It's the 8th inning. Your first priority has to be "get this sucker tied up."