98.65%

[Chichiri]

Here's what I did: I took the Obama prices from Intrade, used those as probabilities of Obama winning that state (for simplicity, assume that McCain's probability is 100-Obama), and performed a Monte Carlo simulation of 5,000 trials per run, similar to what TIA did. I ran the spreadsheet about a dozen times and averaged the results.

Per Intrade, Obama currently has a 98.65% chance of winning the election.


Any questions?

[ddeclue]

1. Be careful in your assumptions...

these sorts of sims tend to assume that there is a simple Gaussian statistical distribution which may or may not be true.


Doug D.

[riverwalker]

2. Hey 98.6

it's good to have you back again
(old sixties song....never mind)

[iceman66]

3. This is of course assuming no election fraud ala Diebold

In a FAIR election, I'd rate Obama's chances about that high. I'm just not sure we're going to get one!

[Realityhack]

4. No it assumes a level of accuracy in intrade. Election fraud...

would be factored into the share price.

[iceman66]

5. Diebold could swing the election to McCain

regardless of the accuracy of the prediction model. It's an independent variable.

[Realityhack]

6. Not on intrade.

Intrade is independent investor estimates. As such it takes into account any fraud those investors expect might happen. Unlike a poll or predictive model.

So NO it is not an independent variable.