“whether a finding is statistically significant is not the same thing as whether your current state of uncertainty is less than it was before or what the economic value of that uncertainty reduction would be.”