The Election Betting Markets Fell Short. They’re Still the Most Flexible Predictor.
The Good: We have divided government. Since Democrats no longer control Congress, they can’t bankrupt America quite so fast!
The Bad: Prediction markets, which I touted as the best guide to elections, didn’t do so well. Yes, they correctly said Republicans would take the House, but they’d also predicted Republicans would take the Senate. Polls and statistical modelers like Nate Silver did a bit better this time. They also said Republicans would win both, but they gave them only a slight edge.
As I write this Wednesday morning, Republicans have (according to ElectionBettingOdds.com, the site Maxim Lott created that tracks election betting around the world) a 19 percent chance of winning the Senate.
Nineteen percent isn’t zero; they could still win the Senate, but Republicans don’t have the 60 to 70 percent chance that bettors gave them in recent weeks.
The Good: Bettors at least adjust their predictions quickly.
Tuesday night, while clods on CBS still said “Democrats and Republicans battle for control of the House,” those of us who follow the betting already knew that Republicans would win the House.
Historically, bettors have a great track record. Across 730 candidate chances we’ve tracked, when something is expected to happen 70 percent of the time, it actually happens about 70 percent of the time.
That’s because people with money on the line try harder than pundits to be right. They also adjust quickly when they see they’ve made a mistake.
At 8:23 p.m., with just 12 percent of the New Hampshire vote counted, bettors gave Democratic Sen. Maggie Hassan more than a 90 percent chance of winning the Senate seat, up from 63 percent earlier in the day. You wouldn’t have noticed that shift watching TV. The AP didn’t call the race until 11:39 p.m.
Bettors also failed to predict President Donald Trump’s win in 2016. But they at least gave him a 20 percent chance, much higher than most “expert” statistical modelers, like the Princeton Election Consortium, which gave him an absurd 0.01 percent cha
Article from Reason.com