The thing you're missing here is that big upsides are low probability, and reduced downsides are high probability. Swing big, and you might be the guy that knocks it out of the park, but you'll probably be the guy that gets struck out, whereas the guy hitting the ground ball might not get a grand slam, but he's got a good shot of getting on base.

I suspect that the expectation value of cutting costs is going to be higher, but with a much lower variance.