[crasher]

I think there is some difference between what we might expect as the "mean" result and what me might expect as the "median" result. I think, even if you purposely downgrade GOP odds in every Senate race by a modest amount just to be safe, you still come up with a calculation for an average GOP loss of under 4 seats. And I bet tradesports caluclations, which look mostly close to right, would come up with a GOP loss right around 3.

But, at the same time, it is probably still fair to set the over/ under higher than that. Probably somewhere between 4 and 5.

[Torie]

The bottom of line of that point of view is that the bell curve tails are longer in the Dem direction than in the the GOP direction, which causes the mean to vary from the median. I agree with that. Good insight. The bell curve is not "normal" as stat folks would say - i.e., it is not symmetrical. It would by useful to distinguish between the two, but how many political pundits understand statistics, and think about mean versus median? :)