It’s too bad that Intrade’s probabilities of winning do not allow the projection of the share of the total vote, so nothing can be inferred about the rankings of those who do not win.
But at this time, that market is making a strong prediction that McCain will win New Hampshire. Romney will be second, a guess based on his second from highest probability.
Meanwhile, Obama continues to strengthen among Democrats and Clinton fade. That is one amazing and swift change, given Clinton’s once large and long-held lead.
A prediction market is hard to beat, since it aggregates all information everywhere from anyone who cares to risk their cash. If you know something it doesn’t, go ahead and hazard some cash. Then you will affect the odds, perhaps. Otherwise, talk is cheap.
But it is predicting something that is hard to predict. It will make mistakes. It will be volatile when new information hits the market. It is very far from being totally useless, however. So far, the predictions have not been simply random or chaotic. The underlying processes may be fat-tailed, or 6-sigma events may occur more than with Gaussian distributions, but the market still works — as it does with stocks.
3:36 pm on January 7, 2008 Email Michael S. Rozeff
