Only replying here to your math (and recognizing it wasn't your main issue) as it's important to remember when looking at run expectancy that the average value is not usually sufficient for making the decision, the spread also really matters. To take something to an extreme, a situation where you're guaranteed exactly one run has exactly the same expectation as a situation where you get zero runs half the time and exactly two runs half the time. I suspect bunting is more likely to get at least one, though less likely to get two or more, while swinging is probably more likely to get both zero and two or more. That is, swinging likely leads to a more variable outcome, even though the average expectation is similar. To build on your comment about the game situation mattering, if runs are scarce, it's easy to see that bunting would be the better play, while if you're down by a couple, swinging away would be the better play. So this isn't a situation where data isn't useful, but it is a situation where you need to make sure you're looking at the right data.