I don't understand - Conrad assured everyone that is was over and Bush was going to win, no question. So why are you bothering?

Ok, that's fascinating. In my head, I sort of thought the swing states included Pennsylvania and Ohio. Your analysis though seems like a great approach. Let the market filter the available information and then decide. Personally, I live in NY. I assume NY will go Kerry no matter who I vote for, although it is likely to be Bush. There may be a lot of people, like me, who could be swayed by the debates.

Also, bear in mind that Colorado could change big time to permit splits of electoral college votes as a percentage of the popular vote. Right now CO is a winner take all state but there is a proposal to change that.

Ok,
Here are the potential problem I have with this particular futures market.

1) Number of actors in the bidding process? Are there enough people participating to make any variance random, or are their few enough that their is a Bush skew?

2) Age of information. In a normal market, traders can expect to receive quarterly reports from companies or daily prices for commodities. Then, futures traders can examine their own predictive capacity with actual results. My concern is that since the last 'reporting day' was four years ago, traders with poor predictive powers are not punished (ie, weede out of the market place). Hence, there seems to be a severe bias in this market towards simply assuming every state that went Bush last time will go Bush this time.

Rusty: Fair points but I think I can deal with them. Firstly these contracts seem reasonably liquid: they have all traded at least a couple of thousand times. If they're weren't enough people trading it would open up opportunities to make money by trading against the biases of the traders currently in the market.

Your second point can be dealt with the same way. Traders may have poor predictive powers, but so do opinion polls. As I said I';d rather trust people investing cash in the result rather than someone answering a market researcher over the phone. The traders are "punished" because if they are wrong, they lose hard cold cash. If I were trading this I'd be scanning opinion polls and talking to contacts in each state to find points of value and to trade them. There are people no doubt already doing it. While we can argue about the merits of the efficient market, you have to assume this is pretty close to one.

RP: thanks for the tip on CO.

Where was the first spot that you'd go for information on betting tips? Those same polls that you discounted.

And "losing money" is not equal for all people. Some people would think HK$30 for a cup of coffee everyday would break the budget. Others are bound to hang out in Pacific Coffee in IFC with great regularity.

And sports betting where there are defined "teams" like the Donkeys and the Elephants will create a skew on the betting based upon team preference. If asked to place a small bet on a Rugby 7's match between the Wallabies and the English, isn't there a bias towards being "a homer" and betting on the side that you're cheering for?

Tom:

Where people get their information is irrelevant to the merits of this method. If I buy a share on the sharemarket, it might be because I do research on the company or it might be because I like the name. The point is that people are staking real money on the outcome. Sure for some (probably most) it will be an amount they don't mind losing, but nevertheless it is real money so there is a sense of loss. In an opinion poll, there are no consequences for giving a "wrong" answer. Losing some money, even if it is small amounts, is a punishment for being wrong.

This also answers the problem of "skew". I could be a rabid Republican or Democrat, but would I risk my hard earned cash for the result? Maybe for a few, but it would simply push prices to a point where more neutral observers could make money from the incorrect pricing.

I'm going to search for other odds to see if it correlates with this set.

To validate the method, you could just see what the track record is for this method in predicting the outcome of previous elections. I'm guessing that, since the pundits tend never to consider it, it's probably not much better than a random guess.

I, personally, think that the method is flawed and that the traditional polls are much more valid for an obvious reason: traditional polls ask "Who are you going to vote for?" but the gambling method asks, "Who do you think will win?" It reminds me of an old spoof-article in "The Onion" which began:

WASHINGTON, DC—According to a Gallup public-opinion opinion poll released Monday, a solid 85 percent of the American people strongly believe that the American people no longer strongly believe that Bush is performing effectively as president.

"Due to perceived dissatisfaction over the economy, a strong majority of Americans believe that a strong majority of Americans believe that Bush's reputation has taken a hit," said Paul Mallock, a spokesman for Gallup. "In addition, we discovered a small but growing minority that believes a small but vocal minority is dissatisfied with the way the president is handling the situation in Iraq. The small but growing minority we found believes that a small but vocal group of Americans thinks that reconstruction is messier and more expensive than Bush originally said it would be."

Of the 10,577 U.S. adults polled, 8,891 "strongly agreed" that more Americans "strongly disapproved" of the president's current performance.

The gambling method doesn't address who will win, but instead it addresses public perceptions about who will win - which is an entirely different subject.

My pleasure, Simon. Hope it's helpful.

Also, here is a link to an article from yesterday's NY Times on the market: http://www.nytimes.com/2004/08/08/politics/campaign/08points.html

I thought you might enjoy it.

For those interested, I've posted my latest survey of Electoral College tracking / prediction / projection / forecast sites (including Tradesports) here.

Executive summary: of the 30 sites surveyed, 22 show Kerry winning, and 5 others show him ahead. Three sites show Bush winning. The approximate average of all sites is Kerry 296 - Bush 242.