Presidential Prediction Contest

Modesty forbids me, but honesty compels me: my 15-month-old predictions for the 2008 Presidential elections have thus far been so spot-on, it’s spooky. I know that many of you have clamored for us to drop the science stuff from our blog entirely, and just talk about politics and/or our personal lives, topics that are severely under-served in the blogosphere. My own preference would be to focus exclusively on physics, to the exclusion of any other topic of any possible interest, but who am I, anyway? This is a blog, after all, and I think we can all agree that the loudest commenters should have final say on what we post about.

Therefore, I feel compelled to offer up another round of predictions, now that we’ve narrowed the field to two major candidates. By why not make it more fun and have a prediction contest? Anyone can join in, just by leaving your prediction the comments. Entries that appear before the end of June will officially count.

But to make things somewhat science-y, let’s use equations to judge who will win. Each prediction consists of two numbers: the fraction f of the total popular vote cast for the two major candidates that goes to Barack Obama, but also the standard deviation σ of your prediction for that percentage. We are thus ignoring the electoral college entirely, and dealing with the annoyance of third-party candidates by concentrating exclusively on McCain vs. Obama. And we are assuming for purposes of misleadingly-precise quantification that each prediction follows a normal (Gaussian) distribution:

$latex displaystyle P(x) = frac{1}{sigma sqrt{2pi}} expleft(-frac{(x-f)^2}{2sigma^2}right) ,.$

And here is the rub: the winner is not the one whose fraction f is closest to the final answer, but the one whose value of P(x) is the highest, when x is equal to the fraction of votes Obama actually does win. The smaller your standard deviation is, the higher your P(x) will be for x very close to your predicted value f , but the faster it will die off as you get further away. So if you are extremely confident, you can ensure victory by choosing an appropriately tiny standard deviation on your prediction. Contrariwise, if you choose a large standard deviation, you might get lucky if none of the confident folks comes close to the actual result. Cool, eh?

So here we go: I predict that Obama will win 55.5% of the popular vote fraction, with 1.5% standard deviation. That’s right — a blowout. Might be crazily optimistic of me, but right now the portents are good. In Obama’s favor, the current electoral map is extremely favorable (not that it matters for our contest), he is an energetic and charismatic campaigner, his organization is impressively seasoned and effective, he will have twice as much money to spend, Democratic identification among voters is soaring, the incumbent President is world-historically unpopular, various economic crises are putting the squeeze on middle-class voters, the war in Iraq is hugely unpopular, and McCain is a bumbling and unconvincing candidate with a tattered organization, little support among the party faithful, a disturbing penchant for changing his mind and misunderstanding his own policies, and little interest in anything other than foreign policy. In McCain’s favor, Obama is black and his middle name is Hussein; also, McCain has a great rapport with the press, who respect his maverick image. Overall, I think the scales are pretty heavily tilted on this one, and I will not be surprised if McCain replaces Bob Dole as the Presidential candidate that Republicans would most like to pretend never happened.

Of course, I could be wrong. So let’s hear your predictions! The winner will receive a lifetime subscription to Cosmic Variance. Or maybe a T-shirt, if we get caught in a generous mood.

114 Comments

114 thoughts on “Presidential Prediction Contest”

  1. And sorry, all distributions must be Gaussian; anything else is too much work for the score-keepers.

    But using a Gaussian distribution is so perverse! The outcome will be somewhere between 0 and 1, but any Gaussian has a positive probability on the whole line.

  2. I predict that, if Obama doesn’t win, there will be 4 more years of dark ages in American science. And instead of a gaussian distribution I’ll affirm it with a Dirac’s delta.

    Now for the booking

    Obama 56.1% with sigma 1.8%

  3. Peter de blanc,

    from what I’ve read about american electoral arrangements, a 150% victory for McCain is not impossible…

  4. f=51%, sigma=1.2%
    Obama will most likely carry the popular vote IMHO, but I suspect McCain will manage to win the election quite easily. Am I allowed to play as a non-american?

  5. I’m going for the shocker. McCain’s not going to make it to election day, but will still get some votes. So under the rules, f = 94%, sigma = 3%

  6. I call no fair on the Gaussianmahoosit! I know not of teh Gaussianmahoosit!

    Stupid science. I thought we weren’t going to have to read about that stuff any more?

  7. Fermi-Walker Public Transport

    Agree, all signs are for a blow-out, but I suspect there will be an October surprise or two which may cut it back a bit. So I will say an Obama win with f=53 % and an S.D. of 0.5 %. My T-shirt size is large.

  8. under the rules, f = 94%, sigma = 3%

    Now this is just silly. If your point estimate is that far from the mean, you want your sigma to be way bigger than that. Remember, you’re not trying to maximize the expected value of P(x), you’re trying to maximize the probability that your P(x) exceeds anyone else’s.

    … which my entry clearly fails to do. Late movers’ advantage. Can we change our guesses?

  9. Hmmmm, just saw the note that comments would be closed 7/19. Does that mean all entries since #46 are invalid?

  10. Given what happened to Gore and Kerry, I won’t put it past the US public to vote for my McCain.

    “I don’t wanna have a beer with the guy!”

    “He’s agonna get sworn in on teh Koooooooraaaaaaannnnn!”

    And the suggestions of invasion of Iran or orchestrated terrorist attack sound all too plausible to me by now (what has the world come to?). At the very least I’ll bet a round of beer for everyone that the ‘terror alert’ is gonna keep rising up till the election and be at some ridiculously high level on the day itself.

    So … 40%/0.8%. Please please please prove me wrong.

  11. Thomas W. Swidarski, CEO of Diebold

    f = 48.632%
    sigma = 0.002%

    Look, you might as well give me the t-shirt right now.

  12. Matt (the real one)

    I think Sean is pretty close to right in terms of actual votes cast, but all we have access to is recorded votes cast. And given the republican edge in vote rigging, I’m going to go with:

    Votes recorded for Obama: 51.8%
    Standard deviation: 1.1%

Comments are closed.

Scroll to Top