## Wednesday, September 20, 2006

### the wisdom of crowds says the democrats still have a 43% chance with hillary. the wisdom of jeremy says there is money to be made off this.

Longtime readers know that a certain very close "friend" of the proprietor is a fan of political prediction markets, and in fact used such markets to win a nontrivial amount of money in the 2004 election. Awhile back, this same "friend" sent Tradesports.com an e-mail asking them to add a set of contracts to their available wagers on the 2008 Presidential Election and, kindly, they did.

Before, you could bet on who you think will be the Democratic or Republican nominee and you could bet on which party will win the Presidency, but not on who will win the Presidency. The reason my "friend" wanted them to add the last kind of wager is that with it, you can calculate the market's assessment of the relative strength of Hillary Clinton versus Someone Else as the Democratic Candidate. The relevant current market probabilities from Tradesports.com:

Probability that HC is the Democratic nominee: 42.0%
Probability that a Democrat wins the Presidency: 48.5%
Probability that HC wins the Presidency: 18.1%

So, what these numbers mean is that the market thinks the Democrats have a 43.1% chance of winning the Presidency if Clinton is nominated (18.1/42.0), compared to a 52.4% chance of winning if Clinton is not nominated ([.485-.181]/.58). If you think these numbers do not reflect the reality of the impact of Clinton on the Democrats' chances, you are of course entitled to your opinion. The great thing about prediction markets, though, is if your opinion is correct, you can make money off it.

(Given that I say things like "Hillary can't win," my own opinion is that the Clinton penalty to the Democrats is larger than what these numbers would indicate. My "friend" has not yet bet on my belief, however. I invite you to profit from my acumen, however.)

A+ said...

How is that legal?

jeremy said...

TheInternetDog said...

So the Democrats have a 43.1% chance of winning with Mrs. Clinton, and a 52.4% chance of winning without her. What about the other 4.5%?

jeremy said...

The probabilities are not supposed to sum to anything. It's like saying there is a 20% chance of rain on a given day in April and a 10% chance of rain otherwise. There's not some missing 70% of something out there.

Anonymous said...

Since 'all politics is local', give us your take on the Patrick/Healey race? (Are you following Massachusetts politics?)

Anonymous said...

i think we should be careful with how we use the term 'wisdom'. predictive accuracy is one thing. outcomes are another...

gabriel said...

And with whom exactly would the Democrats have a better chance?

The thing is that its easy to say Clinton has negatives in the abstraction but this becomes a bit harder to say when compared with any realistic alternatives. My intuition is that the Democrats would have much better chances with Clinton (particularly the centrist 2008 Hilary not the 1994 Hilary) than with -- for example -- Kerry, Biden, or Edwards.

jeremy said...

Gabriel: If you don't think the Democrats have a better chance with whoever the alternative candidate would be than with Clinton, then Tradesports offers you the opportunity to profit from this opinion! (For that matter, you could profit off my "friend" if he enters the market, since you would be trying to move the market in opposite directions.)