Thursday, September 28, 2006

Predictions Markets Map of Senate Races

David Pennock, Yiling Chen and I created a Prediction Markets Map of the 2006 US Senates Elections. When you load the map it downloads the latest prices from Tradesports and colors each state appropriately, a mix of red (Republican), blue (democrats) and green (other candidates). We created the map because predictions based on market prices tend to do better than experts or polls. As Wisdom of Crowds author James Surowiecki wrote recently in the New Yorker
In the past few years, a host of prediction markets, as they're usually called, have appeared online, offering people the chance to speculate on subjects ranging from the box-office performance of Hollywood films to the outcome of Presidential elections and the spread of bird flu. These markets' forecasts have proved remarkably accurate—just as bettors collectively do an exceptionally good job of predicting sports results. (In 2004, for instance, Tradesports, a Dublin-based prediction market, called thirty-three out of thirty-four races in the Senate correctly, and called all fifty states correctly in the results for the electoral college.)
Scaling colors according to probabilities is a surprisingly hard problem. It relates to human perception, we want our eyes to distinguish colors related to probabilities that are not that close. For the map I used a transformation based on the XYZ color space that seems to give reasonable though far from perfect results. Even small projects like this can really bring up complex issues from outside the theory world.

15 comments:

  1. Lance -- How do you know you colored the map correctly? I thought you were color-blind :-)

    ReplyDelete
  2. I just plugged this map on the national politics site MyDD -- I hope it doesn't overwhelm you with traffic. I also asked the community there (or the small subset that will look at my diary) whether they prefer polls, expert opinion, or trading markets to predict races like this.

    ReplyDelete
  3. If the dems don't win, the Frist and others will probably render tradesports illegal in the US.

    ReplyDelete
  4. If the dems don't win, the Frist and others will probably render tradesports illegal in the US.

    Its a republican (John Warner, (R) from Virginia) who has recently prevented a ban on online gambling from becoming part of a larger, must-pass bill. The gambling lobby does not really follow party lines and there are a comparable number of self-righteous anti-gambling pols on both sides.

    As for the post itself, I'd like to see an actual quantification on market forces being a better predictor than polls. Senate races are more often than not, safe for the incumbant and the dominant party in the state. Its not all that impressive that it predicted 33/34 races.

    A better way to gauge its success would be for it to predict the *margin* and then quantify how accurate it was/wasn't in that regard.

    ReplyDelete
  5. [a better test would be to predict the margin rather than just the winner of each race]

    The problem with this is that the predictions market is not trying to predict the margin. (At least if it works like IEM.) The probability of X winning on the map is the current market price of a contract that is worth $1 if X wins and nothing if X loses. If you wanted to use a market to predict margin, you would trade "vote share" contracts where the

    ReplyDelete
  6. [whoops, hit "send at the wrong time!]

    ...where the value of the X contract after the election is the percentage of the vote that X got. IEM has experimented with this.

    ReplyDelete
  7. Suppose that everybody used these market prices instead of poll results for predicting outcomes. Doesn't that provide incentive for a politically interested party to put down a lot of money on certain candidates in order to try to influence the outcome? Similar to how the media releasing exit polls early affects the outcome of the vote (i.e. voters see the results of the exit polls and either decide to show up or stay at home, depending on their political persuasion), these parties may use their financial clout to bias the results of the election.

    I know that this is silly, but it's something to consider.

    ReplyDelete
  8. Its a republican (John Warner, (R) from Virginia) who has recently prevented a ban on online gambling from becoming part of a larger, must-pass bill.

    Yes, but I think his objection was more that unrelated items shouldn't be included in a defense appropriations bill. As I recall there were 9 unrelated add-ons to that bill, and he objected more out of principle than any stated desire to keep on-line betting legal.

    Doesn't that provide incentive for a politically interested party to put down a lot of money on certain candidates in order to try to influence the outcome?

    TradeSports is very easy to manipulate if you are willing to throw a modest amount of money at it because the volume of trading is very small. I know of at least one case where some wealthy individuals did throw some money at a political wager there for the purpose of manipulating media reporting of the topic.

    ReplyDelete
  9. What does it mean when the sum of probabilities for a state is more than 100%?

    ReplyDelete
  10. Many states have low trading volume with a wide bid-ask spread so the last prices in the individual securities don't always add up to exactly 100.

    ReplyDelete
  11. [what does it mean when the sum is more than 100%?]

    It means that there is an arbitrage opportunity that has not yet been exploited. In the IEM market, if the GOP-win and GOP-lose contract prices totaled $1.04, you could in principle buy bundles of one contract each for a total of $1.00 each and then make money selling them to the people offering the current prices, until the price went down. But there are stickinesses in the markets -- the price quoted is not the "bid" or "ask" price but some sort of average of previous trades.

    There's an interesting and heated discussion of the validity of these markets as a tool for predicting elections, currently going on at MyDD.

    ReplyDelete
  12. We all know coloring is hard. Using stripes or dots with the spacing indicating margin of advantage would be more user-friendly.

    ReplyDelete
  13. The map is nice. The map makers' comment is hypocritical. I state why here: http://caveatbettor.blogspot.com/2006/10/you-call-it-corn-i-call-it-amazing.html

    ReplyDelete
  14. What about a similar map for Governor races?

    ReplyDelete