Politics vs. the playoffs: Why do predictions change over time?

David Rothschild
The Ticket

A lot of people have asked me about the volatility in my predictions. What can we learn from movement in the prediction markets? With the current election cycle only getting started, I am going to provide a few examples from another type of campaign, which is much closer to its finish. All of the below likelihoods are derived mainly as an average of the prices in the major prediction markets; I've explained the process in more detail here.

On Opening Day of the baseball season, I gave the Phillies at 19.2 percent likelihood to win the World Series. I now have them at 32.3 percent likelihood. The Phillies season has gone just as expected, with a few bumps along the way. Why has the likelihood changed? Because the march of time has eliminated the uncertainty that a long campaign brings to the outcome. As the days before an event become fewer (whether the playoffs or Election Day), the leading contender has a higher likelihood of victory, because there are fewer unknowns that may shift the course of the outcome:

Less than a month ago, the Indianapolis Colts looked like a strong contender to win the NFL's Super Bowl. Then the news of Peyton Manning's injury unfolded, and their likelihood of winning dissipated to near zero. In this situation the movement actually quantifies the value of the missing player (or, in politics, the cost of a campaign mistake). In the case of the Colts, it is clear that the market assumes that nearly 100 percent of the team's Super Bowl chances are tied to Manning's health:

As the local, state and national campaigns for elective office in 2012 play out over the next 15 and a half months, expect similar charts from me. A relatively uneventful campaign for the favorite would look very similar to the Phillies' chart, while an unpredictable "Macaca moment" for an otherwise strong candidate would look a lot like the Colts when they lost Manning.

In case you are wondering, here are the likelihoods of winning the World Series for the eight teams (their leagues are abbreviated NL for National League and AL for American League) that would make the playoffs if the season ended today and thus the most likely eight teams to win the World Series: Philadelphia Phillies (NL) 32.3 percent, New York Yankees (AL) 17.1 percent, Boston Red Sox (AL) 11.4 percent, Detroit Tigers (AL) 8.2 percent, Milwaukee Brewers (NL) 8.2 percent, Texas Rangers (AL) 7.1 percent, Atlanta Braves (NL) 5.5 percent, and Arizona Diamondbacks (NL) 4.5 percent.

Here are the likelihoods of winning the Super Bowl for the top eight teams in the NFL: New England Patriots (AFC) 14.7 percent, Green Bay Packers (NFC) 13.1 percent, Philadelphia Eagles (NFC) 10.9 percent, San Diego Charges (AFC) 8.1 percent, Baltimore Ravens (AFC) 7.8 percent, Houston Texans (AFC) 5.7 percent, New Orleans Saints (NFC) 5.5 percent, and New York Jets (AFC) 5.5 percent.

David Rothschild is an economist at Yahoo! Research. He has a PhD in applied economics from the Wharton School of Business at the University of Pennsylvania. His dissertation is in creating aggregated forecasts from individual-level information. Follow me on Twitter @DavMicRot and email me at PredictionBlogger@Yahoo.com. In the interest of full discloser, he is a proud Phillies season ticket holder.