I would be interested to see a statistical analysis showing the benefit of a marginally faster runner at second with 2 outs vs. losing potential extra inning at bats from your best hitter. At face value, I never like it, but I don't doubt that it may be the statistically right thing to do.