Intrade is pretty cool
 

I've known there were places you could bet on ridiculous stuff like whether or not Brad and Angelina will get a divorce, but it's pretty cool getting into the details of this.

There's not a ton of money changing hands over all the political primaries, but just like I'm a big believer in efficient markets, I have to believe intrade is probably the best predictor you can find on the election.

Anyone actually trade on intrade? With the relatively small amounts of money traded, I would think though it's still a good predictor, there are some areas to exploit, if you felt like you had good insight on something.

There is a growing literature on the intrade contracts in economics. You are right to worry about the depth of the market. For a lot of contracts the market is pretty thinly traded. However, it is going to be by far your best predictor.

For example, by early Tuesday the intrade market swung toward Romney in a big way. I think the best explanation is that some traders clearly had exit poll information early before it was made public:

Here is the link to the thread were I pointed out the change in prices in favor of Romney on Tuesday afternoon:

http://cougarguard.com/forum/showthread.php?t=15926

Yes that was the post that piqued my interest.

Pelagius, what do you think of this strategy?

Take the safe bets that have between 50% and 80% chance.

for example,
Hillary for Dem candidate .58
all three McCain, Guliani, Romney for Rep candidate .80
Hillary Florida .84
Romney in Nevada .85
McCain South Carolina .62
both McCain and Guliani in Florida totals .75

You don't win much after commission, but you keep churning those. Makes me wonder if you could exploit it. Maybe there's a bias towards the crazy bet?

It's possible that candidates with little possibility of winning are trading at too high of prices. There is some evidence in the literature that investors like stocks with positive skewness (the possibility of a big payoff) and hence are willing to pay "too much" for them. But I wouldn't call the evidence overwhelming.

Another thing to do is arbitrage between the Iowa and intrade market. The problem is, of course, the Iowa market. You can't place big bets so you will probably only be able to make like $5 an hour. But there are sometimes/often differences in prices that are big enough to exploit after taking into account transaction costs.