This is rather off topic, but one thing thats always bothered me about reliever usage is when there is a reliever on the mound who has been effective, hasn’t pitched for long and is taken out for a situational advantage. It seems to me, pitchers frequently either have their stuff or don’t on a particular day. If you have a pitcher in who is obviously throwing well, why not leave him in instead of risking bringing in another pitcher who may or may not be having a good day.

I’d be curious to see some sort of analysis of play by play data to see what the expected outcomes are for a pitcher coming in cold, vs a pitcher that had gotten one out previously, vs pitchers who had gotten 2 consecutive outs, etc. If the BA for hitters against pitchers who had gotten 2 outs without giving up a hit is significantly lower it would seem to be an advantage to leave that pitcher in rather than trying to get some smaller advantage out of a platoon split.

Another case would be if the reliever had pitched the previous inning scorelessly is he more likely to pitch well the next inning. That would probably be easier to check. I haven’t read too much performance analysis stuff, so somebody might have done this already, but I’d be curious about the results…