Remember, remember the 5th of November

Ron Paul has now passed Fred Thompson in the probability of winning the Republican nomination.  According to Intrade, Paul has a probability of winning the nomination of 8.8%. (Guiliani (42.0%) and Romney (27.6%) are first and second.)

In closely related news, Paul raised $4.2 million yesterdayV.

Thanks to Barry Klein and Tim Groseclose for the tips.


Hi there. Alex, there may be a mistake in this --people make many mistakes in poll reading by translating statistics into winning probabilities. If a poll shows someone ahead of another by 10 percentage points, it does not mean that he has only 10% more chances of winning. if Candidate A has 60% of the poll, I may bet that he has 93% of winning (depends on variance). Now this may apply to polls about who is going to win (as compared to who are you going to vote for).

Intrade is a prediction markets site, not a polling site. If you think that the numbers there are mistaken, cash in.

Does anyone know if Paul is playing this thing to win or is an "issues candidate". You gotta play the primaries to win, does he have an ground game in the early states?

"Alex, there may be a mistake in this --people make many mistakes in poll reading by translating statistics into winning probabilities."

Exactly how stupid are you accusing Alex of being?

GoodneesOfFit: Ron Paul is running to win. He has been very clear on that, though it's taking some people some time to figure that out. He does have some ground in Iowa and NH, something like 7% in NH right before he spent money on tv and radio ads. It will be interesting to follow.

closing the tag.

does he have an ground game in the early states

Yes. He has Drew Ivers as campaign chairman in Iowa (guy has run a lot of campaigns before); he has a bunch of meetup groups in both Iowa and NH. There's a push to try and get a thousand volunteers to NH in the days and weeks leading up to the primary. He's also been running radio and now TV ads in both states (does that count as a ground game)?
I have high hopes for both Iowa and NH. In the former state, the low voter turnout (6% or so) typical of the caucuses favors a highly motivated core of supporters. In the latter state Paul should have as receptive an audience as any he is likely to find in the US, and the primary is an open one.

The problem is as follows: I estimate that to have a 8.8 pct probability of winning he needs to be between 16 and 40% in the polls. Are his polls so high?

I live in the relatively early state of Nevada, and there are *far* more Ron Paul signs up around Las Vegas than for *any* other candidate.

Any stats on the past accuracy of intrade predictions? (i.e. 2006 elections)

...I'm pretty sure that somewhere in Proverbs, it says (in the King James):

Talkest not about what thou knowest not, especially on topical blogs, lest thine ass be shown.

Polls are like balance sheets, they only capture the present (and dinosaur land line phone polls can be as misleading as Enron balance sheets).
Quotes from Intrade are of binary options. Options capture both the present and the future.

The market is not exactly liquid. Let's consider the possibility that Paul's Internet-savvy supporters are bidding up his Intrade contracts to make him appear to important people far more likely to win than could possibly be the case (SERIOUSLY). (Or the closely related scenario that Paul's Intrade-aware supporters are buying the contracts in the self-deceptive optimism that he has a chance). That's a lot cheaper than a few TV or radio ads and it packs a lot more punch, among nerds at least. To sell these contracts you'd have to wait something like 6 months to get a ~8% ROI: i.e., barely worth it.

I say let's bid it up even more...slowly.

Let's consider the possibility that Rudy's Internet-savvy supporters are bidding up his Intrade contract and shorting Paul's contract below it's true market value.

By Paul N, you mean Sean Hannity right? The guy who is so paranoid that Paul's supporters are voting multiple times in polls and the other candidate's supporters only vote once.

NNT, how did you calculate those required poll numbers?

Comments for this post are closed