Skewing my winnings

By Razib Khan | September 29, 2012 12:42 am

Last spring I made a bet with a friend that Mitt Romney would win. He gave me 5:1 odds, and I assumed a 40% chance that Romney would win. So I expected to lose, but if I won I’d win big. At this point I assume I’m out that money, because I’d put Romney’s chances at less than 40% (though I think people underweight uncertainty, so I believe there’s a lot of variation in this prediction). But now I’m hearing/reading that many Republicans are under the impression that the polls are skewed. If you believe that the polls are skewed would you be willing to bet money that the polls are skewed? Specifically, I want to wager that “unskewed polls” turn out to be further off the mark than the regular polls in reference to the final election results. I’m not 100% sure that the pollsters are correct, and I don’t know more than a superficial amount as to the weighting methodologies, but the track record of skew-skeptics is suspect enough that I think this is a way I can make money off people who I perceive to be suckers. Of course, the people who I perceive to be suckers think I’m the sucker, which is fair enough. Take my money! Please speak up in the comments if you want to make a bet. I’d want it to be public, and you have to put your real name out there. Also, I want to know if you’ll give me some odds, because I assume you are moderately confident in your assessment that the polls are skewed.

MORE ABOUT: Politics

Comments (25)

  1. RRaccoon

    No coincidence that the supporters of the losing candidate think the polls are off. Good luck finding any action on your bet.

  2. Rick Cendo

    The people who believe the polls are skewed also believe that the vote is corrupt. So they won’t accept actual results as an intrusion on their alternative reality.

  3. Mr. Bisuit

    How many right-wingers with money do you think actually read this blog….? If you truly want betting action, take your bet to FOX.

  4. Dwight E. Howell

    You might want to go back and look at how accurate polls have been at predicting elections in the past. The track record isn’t great. Even the exit polls in WI were wrong. It appears the Democrats who wanted the governor out stopped and chatted and the people who voted to keep him largely walked on by including a significant number of Democrats who had to have voted for him.

    There is also the non trivial question of how many of the various sub groups are actually going to show up on election day. If you assume that blacks will turn out in the same numbers as his last election you get one result. If you note that the black community has not fared well during his tenure in office and he has deeply offended many black Christians you have to wonder if some of these people are going to bother to show up and vote for him. The Jews have to know his position relating to the Jewish state, etc. He pretty much had a solid Catholic vote last time but he’s at war with the Catholic church. What does this all mean? You’ll find out after the votes are counted.

  5. ackbark

    2. So, you’ll take the bet, then?

  6. RRaccoon

    The polls are for the most part accurate and those who want to elect Romney are simply going to the well of a main stream media that is out to get them. Although Obama has alienated many ( if he didn’t he wouldn’t be governing), Romney’s main opponent appears to be himself.

    If the bet is to take the side of polls being off and your side is their accuracy, you will win. The trailing team is of course going to try to give his supporters hope.

  7. wilzard – strange, huh, that if you reweigh the polls to assume more Romney voters than Obama voters, Romney is winning?

    I prefer to stick with Nate Silver’s melded polls, which have a proven track record of being accurate.

  8. #5, exactly. shut up dwight and put up some money. seriously.

  9. The trailing team is of course going to try to give his supporters hope.

    yeah, of course. i myself lean toward romney, as long time readers are probably aware of (and my friends in “real life” are quite conscious of 😉 but that doesn’t mean i’m going to put down a stream of bullshit like dwight above. that being said, if dwight et al. are more than full of shit (i.e., they believe the crap they’re putting out there), then i want to make some money off their delusion. it’s a public service, incentivizing people to be more attuned to what is, and not what they want to be.

    we’ve seen this before

  10. dwight, i want some odds and parameters of your model confidence. then i will throw out some monetary numbers. let’s get this done! 2012 election may still be a financial boon for me 🙂

  11. dave chamberlin

    I trust Nate Silver of the fivethirtyeight blog and the New York Times. He now has the odds right around 78% that Obama wins, right about your 5 to 1 odds. If somene wants to cut through the crap that passes as political analysis but is just TV talking head bullshit seek out my main man Nate Silver.

    Anyway I’ll take your bet, not for the money but for the fun of it. If I lose I’ll mail you five books I would hope you haven’t read but would at least enjoy skimming through. If you lose you mail me one book. The loser e mails the other his home address.

    I think Romney loses by more that the six point he is trailing by. The man has no game. It is perfectly fine to be a rich asshole and become president, some of our better modern presidents were exactly that. But good ole Mitty campaigns as one. Game on.

  12. The Kutra

    To play devil’s advocate, it is plausible that the polls are skewed, and that fence sitters will switch sides based on who they perceive as winning, making the polls a self fulfilling prophecy.

    Of course, I don’t think this model is likely to be true.

  13. #11, game on! also, anyone who agrees with dwight, but doesn’t put down any money, i’m not publishing your bullshit. (happened a few times) in fact you don’t have to put down that much money, i’d be happy for many suckers to put in small bets so i can make some money.

  14. Chris_T_T

    I remember Democrats making similar arguments ahead of the 2004 election.

  15. Daniel Gonzalez Buitrago

    I´d like to participate in your bet Razib. I´ll bet 40 USD on the following:

    – The latest poll by Pew research before the election will overestimate Obama´s share of the national vote by at least 2%. (20 USD)

    – The latest poll by “We ask america” in Wisconsin will overestimate Obama´s share of the vote by an amount greater than its margin of error. (20 USD)


  16. #15, awesome! let me double-check those specific polls to make sure i’m not being a sucker (i was thinking of a polling average like for house effects i’m ignorant of. but props for stepping and being more than a bullshitter.

  17. dave chamberlin

    Nate Silver now has Obama’s chance of winning at 85%. Chances are the race will tighten, that is just regression to the mean, but what section of the population is going to push Obama over the top? Waitress moms, it seems the blue collar women don’t care for Romney and they are going to put Obama over the top in swing states like Ohio. Come on people, start gambling, take some of this guys money.

  18. Karl Zimmerman

    17 –

    I’m not sure why you say “chances are the race will tighten” U.S. Presidents who win re-election basically always end up winning a higher percentage of the popular votef or their second term. I can think of only three exceptions.

    1. James Madison in 1812. The Federalist-supported candidate, DeWitt Clinton, was actually a dissident Democratic-Republican, and in the early days of the Republic the franchise was so limited that in some ways the inside politics of the American elite probably played a larger role in the outcome than any popular sentiment on Madison’s performance.

    2. Andrew Jackson, who had an Anti-Masonic challenger (Amos Ellmaker) in 1832. Discounting this, his share of the two party vote rose from 56.2% to 59.2%

    3. Grover Cleveland, who also had a third party challenger (James B. Weaver – Populist) when he ran for his second winning term in 1892. Discounting the populist share, he improved from 50.1% of the vote to 51.8% of the vote. Plus, of course he famously won the only non-consecutive second term in U.S. history.

    Thus, it’s fair to say that no U.S. president since 1812 has seen their share of the two-party popular vote decline when successfully reelected to their second term. Going by this I’ve always assumed that if Obama looked reasonably good coming out of the conventions, he’d do better than 2008. He very well might do worse in the electoral college however, due to reapportionment shifting electoral votes to red states, plus almost certainly not winning Indiana again. But I’m willing to stake money (or at least a book or two) that he’ll perform better in 2012.

  19. i don’t think regression to the mean is the issue. rather, there is deep uncertainty which can’t be reduced to a number. e.g., a large effect foreign policy debacle. the 85% value is the best you can come up with by taking into account what you know now.

  20. Karl Zimmerman

    Razib –

    You are of course right that there needs to be some uncertainty built into the model. But Nate Silver’s model already has that in there. If numbers remained unchanged between now and election day, the percent chance of winning would continue to rise (until it equaled the “now-cast”) as the likelihood of such “black swan” events would become less and less in the remaining period before voting.

  21. as the likelihood of such “black swan” events would become less and less in the remaining period before voting

    silver can’t account for the magnitude effect of the swan. i’m 99% sure he’d agree with me on this, though i haven’t gotten to reading his book. this is of the same class of problems as formal models in finance. you can account for what you can account for, even uncertainty you can account for. but you can’t formally account for what you can’t account for. this is not much of an issue in physics which can be modeled in a linear fashion. much more of an issue in social science.

  22. ackbark

    In addition to the possibility of black-swan events happening, won’t the scale of the effect of such an event decrease increasingly as the time period between now and election day decreases, e.g. less time for fallout?

  23. #22, my overall point is that these models can’t account for the unaccounted. in physics this is less of an issue. in quantitative social science it is much more of an issue. any sort of assertion of the form you make above has implicitly within it the “all things equal” axiom. an important and useful one, but still a prior.

  24. after checking, i accept #15’s bet. is anyone else more than full of shit?

  25. hey skew-suckers, why won’t you bet me money? you people are full of such crap.* here i thought i was going to go into the black 🙁

    * at this point, you will start writing a long comment, but refuse to bet me any money, because even you don’t believe in your crap.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar