Is Tradesports accurate?

So we used a few numbers from Tradesports before the election to get a feel for how the "market" was assessing outcomes (eg see here, and here).

Tradesports works differently in that it goes up or down according to how many people are buying shares in particular outcomes (such as Dems taking the House or the Senate). It's not magic, because where are people getting information from upon which to make their purchases? Well, polling data of course just like the rest of us.

The difference is that in theory Tradesports reflects what people prefer, as a whole. I'm not sure that that's correct though. When real money is involved, even if you prefer a GOP victory, but think that the Dems will win, which way do you bet? I would think you'd go with your expectation, not your preference.

Do people therefore hold unrealistic expectations? That's really the question for market-based numbers.

Atrios rightly pans Tradesports for not being "magical" and posts this image showing how it blew the Senate call.



I would argue that this shows that the market reflects what people think will happen (ie it's just a reflection of polling) not what they wish or desire to happen. Even Atrios admits he didn't expect the Senate to go Dem.

Therefore it's really just another form of polling.