It seems more to me like you jump up my butt as soon as you see my name without bothering to think, consider, or make any sense whatsoever—then when it’s apparent even to the likes of you that you have made a wildly poor decision, you rinse and repeat in another thread. And then sometimes Scott disagrees with you (“schools” you, perhaps) and you then, thankfully, have very little to say.

As for unhinged arguments, you’d have a Hall of Fame without Randy Johnson. Good lord.

by assuming that the quality of opposing pitching was essentially random. This is probably wrong, as we seem to match Guthrie against opposing teams aces fairly regularly, but it’s a much better assumption than assuming that on a given day the quality of your team’s pitching is random, as it really matters whether Kazmir or Jackson are pitching, and those pitchers rotate. However, a team faces many different starting pitchers of varying qualities, and we can more safely treat the quality of opposing pitching as a random variable over time (again, the major problem with this is that teams all play their best pitcher first, second pitcher second and so on, so it’s not clear we can get away with this).

The statistical concept I’ve been dancing around is stationarity. Pyth records are supposed to be a predictor of how many wins a team gets based on their runs scored and runs allowed, and then the error in the prediction is usually attributed to “luck.” My argument is that such a model may be skewed, because one typical assumption to do these kinds of predictions is that the error terms are “stationary” meaning that the probability of a win would be equal on any given day. This is simply wrong, as the Rays are simply more likely to win games when Kazmir is pitching then win Ediwn Jackson is pitching.