I'm not sure you're aware how high the probability is that Hernandez has better results than Stroman over a 20-inning sample. It's much higher than 0. And those 20 innings don't move the needle at all on the probability that Hernandez will be better than Stroman over the next 20 inning sample, which is why you wouldn't keep Hernandez over Stroman regardless of what happens over the next three weeks.