Election prediction round-up

[libertango]

With one week to go before the event, here's the state of play in the election as of today, according to simulators and markets. Comparisons are to my previous roundup, on October 8th:

Princeton Election Consortium: Obama 363 electoral votes, McCain 175 (Formerly Obama 353 electoral votes, McCain 185, so Obama +10)

FiveThirtyEight.com: Obama 348 electoral votes, McCain 190 (Formerly Obama 347 electoral votes, McCain 191, so Obama +1)

Electoral-Vote.com: Obama 364 electoral votes, McCain 157, Ties 17 (Formerly Obama 349 electoral votes, McCain 174, Ties 15, so Obama +15)

Iowa Electronic Markets: Dem .86, Rep .14 (Formerly Dem .81, Rep .19, so Obama +.05) (one way to think of that is as percentages, but it represents the price of contracts out of a $1.00)

Intrade does both:
* Obama 364 electoral votes, McCain 174 (Formerly Obama 338, McCain 200, so Obama +26)
* Obama .88, McCain .12 in the contracts (Formerly Obama .71, McCain .29, so Obama +.17)

*^*^*

At this point, the changes are flat enough the question becomes, Are the polls accurate? (The one exception has been Intrade going from being a laggard to joining the consensus.)

Everyone else has noticed this, too:

Nate Silver of FiveThirtyEight.com, today: "Stop me if you've heard this one before."

Sam Wang of Princeton on 22 October:"In which I write of paint continuing to dry. There's just so many posts like this a guy can write. Today, Obama is still crushing McCain. Still. Crushing. McCain."

Perhaps the greatest irony: The numbers stopped drifting after the second debate, in McCain's much favored Town Hall format. The electorate saw him in that format, found him wanting, and haven't seen anything to persuade them otherwise since.

Thus, be careful what you ask for.

Comments

Intrade

[hal-obrien.livejournal.com]

Reading through the report in CQ that Krugman cites, it seems the manipulation was to bid McCain up.

Since McCain's contracts have now fallen at Intrade to roughly match IEM, that implies such manipulation is now a) over, b) unsuccessful, or c) both.

It could be that when I sampled Intrade last time around the manipulation was in full swing, and the removal of it has caused the reversion to the consensus I've noted.

Re: Intrade

[holyoutlaw.livejournal.com]

It could be that when I sampled Intrade last time around the manipulation was in full swing, and the removal of it has caused the reversion to the consensus I've noted.

That's what I was referring to, although not clearly.