Sep

28

 I'll go with the prediction market which reduced the probability of wining to 0.30 from 0.37 for Trump. Having listened to the debate, it is hard to believe that anyone could have thought Trump won the debate. He was totally unprepared. He could only mutter "wrong" and talk about his abilities in response to totally prepared remarks from Hillary. The Times gave a good fair analysis of the debate. As did Washington Post and Fortune.

anonymous writes: 

The prediction market is not anything like the markets that you all trade in every day. (Those of us out here in the piney woods are guilty of buying and selling stocks on occasion but comparing what we do to what you all do is like comparing subsistence farming to Cargill.) For one thing, there is no size. The move down 7 points came on basically no volume. If a stock moves that much on that little money changing hands, the shares belong to a business that is already bankrupt. Oh, wait, that is the general business magazine trade these days….and the newspaper business. For those who want to know where the money went…  There is not much volume; but there is, as there used to be with Berkshire in the good old days when its owner did outperform the market, some considerable size.

Victor Niederhoffer replies: 

I believe that the many small traders with insights and incentive to make money keep the prediction market very accurate. And there are numerous studies that show that the prediction markets work much better than the polls. (I think). 

Andy Aiken writes: 

A problem with most prediction models, including prediction markets, for the US presidential election, is that they combine state polls using a Gaussian copula-like approach. Andrew Gelman, political scientist at Columbia, showed how this approach is subject to the ecological fallacy. He resolves this with a hierarchical Bayesian model, which Nate Silver adapted for his 538 site. This is not to say that 538 is accurate, since it is only as good as the polls that are used as inputs, but it avoids a common methodological flaw. It doesn't take too much capital to manipulate a prediction market. I personally know some people close to Bush who in 2004 would push up or down the Presidential Election contracts market on the Iowa Electronic Market, merely to influence perceptions. Brexit may have been an example of such a manipulation attempt.


Comments

Name

Email

Website

Speak your mind

1 Comment so far

  1. Arch Stanton on September 28, 2016 4:11 pm

    Oddschecker provides a nice one-page summary of the current betting lines, with shortening and drifting odds highlighted in color. http://bit.ly/1jey0nm

Archives

Resources & Links

Search