Explanation of Predictions Markets Maps for
2006 Senate and Governor Races

Tradesports.com runs a real-money security market for the democratic, republican and other candidates for each state that has a governor and senate race in 2006, for example MI.GOVERN06.GOP. If the republicans (GOP) wins the governor's seat in Michigan then this security will pay off $100 otherwise it will payoff nothing. Traders can buy, sell, even sell short such securities. The price reflects a consensus value of the probability that the event will occur. If the price of MI.GOVERN06.GOP is 48.3 that reflects a 48.3% chance that the republicans will win the seat.

When you load the map, the last sale prices from Tradesports are converted into colors. The state is colored Red (Republican), Blue (Democrat) or Green (Other Candidates) depending on the maximum price with the shade indicating the amount of the lead. Near white represents a close race. A state colored grey has no race in 2006.

Put your mouse over a state to see the probabilities based on the normalized security prices for that state. Clicking on that state takes you to a graph of historical trading information for the highest priced security for that state. The probabilities don't always add up to 100% because of rounding error.

Below the maps you can find the expected number of govenors or senators for each party as well as the number of races "leaning" towards each party. The expected value has better mathematical properties but we added the leaning values for comparison. Both the expected and leaning numbers includes those in office not running for reelection.

The senate map has probabilities that the Republicans will retain control of the Senate and the House respectively. Clicking on those will take you to the historical information for the corresponding securities.

Electoral-Vote.com uses poll data to color their map. We built this map in response since we believe that market prices give a truer reflection of how each state will vote. Market prices did much better than poll results in the 2004 Senate and Presidential elections.

For more information about the predictive power of prediction markets see Artificialmarkets.com and Chris Masse's extensive collection of prediction market resources.