[Longbow1969]

But I remember way back in my statistics courses one of the first things we learned was that you can never take two polls, or more, and average them.

The RCP average was within 1 point of the final result in both 2004 and 2008. It's developing a pretty good track record.

[Fightin Whitey]

Sure can’t argue you with you there.

But even leaving aside actual rules of probability (which I don’t remember any of) it’s just instinctively hard for me to trust the average of a bunch of half-ass polls, you know?