Tradesport predictions for 2006 elections
The Tradesports and the 2006 election?
Can predictions markets be right too often?
David Pennock
Prediction map post mortem.
Robert Forsythe
We have discussed the use of betting markets to predict the outcomes of elections several times in Chance News. See for example Chance News 12.02
Lance Fornow Computer Scientist at the University of Chicago, David Pennock and Chen Yiling Research Scientists at Yahoo have carried out research to evaluate the ability of markets such as Tradesports, the Iowa Political Markets and other such markets to predict the outcomes of elections, sports events, Oscar winners etc.
In their discussion of the predictions relating to the 2006 Senate races, they provided the following map showing results as of about 9 AM CST election day.
In his analysis, Fornow wrote:
Every state colored blue was won by a democrat and every state colored red went to a republican. But also note the 69% given to GOP (Republican) Senate control although this election will give control to the democrats. No outcome would have made all the states and senate control agree with the 9 AM map.
Were the markets inconsistent? No, because the markets predict not absolutely but probabilistically. For example, the markets gave a probability of winning 60% for each of Virginia and Missouri and the democrats needed both to take the senate. If these races were independent events, the probability that the democrats take both is 36% or a 64% chance of GOP senate control assuming no other surprises.
Of course the races were not independent events and there are other states involved making it more difficult to compare the probabilities of the individual races with that of senate control.
So how did the markets do as predictors? Quite well as the outcome seems quite reasonable given the markets. Other outcomes would have also been reasonable such as the Democrats losing Virginia and the senate remaining in republican hands, a possibility that came very close to happening.
While it seems likely that the outcomes of individual state elections are not independent it is interesting to check this by looking at the data.
To do this we need to describe how the Tradesports betting works. We illustrate how this in terms of betting that Hillery Clinton will be nominated as the democratic candidate for the 2008 presidential election.
Betting at Tradesports is made through contracts. You can buy or sell contracts. In this example the a contract has a price between 0 and 100. A contract with price 0 is worthless. A contract with price 100 is worth $10.
Until the nomination is determined you can offer to buy or cell contracts at specific prices. If your offer to buy matches another persons offer to sell you can bid for a contract. Tradesports makes the initial price for a contract and after this it is determined by offers to buy and sell them. Thus Tradesports is simply a match maker though there are some fees attached to this which are explained here
As of December 23 Hillery contracts were given in the following table.
.
At this time the price of a contract is 50.3. This is interpreted to mean that Hillery has a 53% chance of being nomiated.
The closing price of a contract varies through time as illustrated by the following graphic:
Here are similar graphs for other leading cadidates for democratic and republican nominations as of Dec. 5, 2006 that appeared in a New York Times article, '08 Candidates in a Virtual Market, Dec. 5, 2006.
http://graphics8.nytimes.com/images/2006/12/05/us/politics/1205-nat-webACTION.gif
Now we return to the 2006 election and consider the 33 Senate races. The democrats had to win 24 or more of these races th control the Senate. We assume that the indpendent candidates Lieberman in Connecticut and Bernie Sanders in Vermont will win their races and vote as democrates. If so democrates would have to win 22 or more of the remaining 31 races.
We assume that for a given state and date the current price of a contract that a candidate will win the election is a reasonable estimate for the probability the candidate will win the election. If winning the elections in different states were independed we could then calculate probability that the democrates would win 22 or more of the races. However, as we have remarked, it seems unlikely that they are independent. However the number of states with price greater than 1/2 of winning is a reasonable estimate for the expected number of races the democrats will win if the election were on this date.
Tradesports makes their data available so we can make these caculations. Doing this using the closing price for each day from Jan 18 to November 6 we obtain the following graph:
On any given date we can estimate the expected number of states the democrats will win sd the number of states with closing price great than 1/2.