“Compare to intrade.com, which predicts Republicans gain between 55-60 seats, and yet only 85% chance of Republicans taking the House.”
This makes sense, because an outlyer “october surprise” might happen, so there is that 10-15% chance of a shift that denies GOP a House majority, but almost certainty that we are on track to a big GOP victory.
A Monti Carlo simulation will assume, by default, that the races are statistically independent random events. It will assume that there are no covariances, that there is no systematic relationship between one race and another. There is no physics in it.
This is one BIG reason why this thread is utter and complete crap. Crap.
Monti Carlo simulations are why Wall Street badly invested so much money in mortgages. They never properly figured out the systematic way the price of one house could vary in the same way as the price of another house 100 miles away.
KHR should be banned from FreeRepublic for senselessly pimping its blog.