tag:blogger.com,1999:blog-35364652.post8642172194183091418..comments2024-03-22T10:29:59.088-04:00Comments on Quantitative Trading: Are political futures markets really predictive?Ernie Chanhttp://www.blogger.com/profile/02747099358519893177noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-35364652.post-70080283429413858752006-11-24T20:33:00.000-05:002006-11-24T20:33:00.000-05:00Hi Steve,
One can certainly argue that at any poi...Hi Steve, <br />One can certainly argue that at any point in time, the market consensus is the best predictor of the future given the current information, in the sense that it is better than random guesses. The only problem I have is that news has so much impact on the prediction that any advantage over random guesses is not meaningful in a practical sense. This is different from your example of football pool tallys, where I bet that the prediction is fairly stable over time and not subject to daily news releases.<br /><br />I have not read Surowiecki and so am not sure what he means by applying the technique to the stock market. I thought the stock market is already an implementation of the polling mechanism, and it is the best possible predictor of the future of a company that is available with public information?Ernie Chanhttps://www.blogger.com/profile/02747099358519893177noreply@blogger.comtag:blogger.com,1999:blog-35364652.post-91225286678391552152006-11-24T11:49:00.000-05:002006-11-24T11:49:00.000-05:00Hi Ernie,
There's a book out entitled "The Wisdom ...Hi Ernie,<br />There's a book out entitled "The Wisdom of Crowds" by James Surowiecki which purports to use the aggregated polled results of the public to make informed decisions. Normally, this type of analysis would be used in the stock market in a contrarian fashion, ala Charles Mackay's classic book "The Crowd: Extraordinary Popular Delusions". However, Surowiecki steps the reader through history and shows that polling has positive predictive power when three elements are present: 1)Diversification 2)Independence 3)Proper aggregation of the results.<br /><br />Here is a summary of the technique:<br />In a guessing game where participants are asked to guess how many marbles are in a jar, one would assume that the guesses would be distributed randomly and that the mean of the guesses would be somewhat close to the median. Surowiecki argues that when the three elements are present, the mean of the guesses will approximate the actual number of marbles in the jar and beat the majority of individual guesses (median). <br />I tried the technique by asking the head of the Math Dept. at my daughter's school to get the school wide football pool tallys. After aggregating the individual picks, I found that the mean of the student guesses outperformed 95% of the individual student picks. <br /><br />Surowiecki argues that the technique is notoriously difficult to apply to the stock market. Any thoughts on applicability?<br /><br />Steve HalpernAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-35364652.post-18265176009966631802006-11-17T16:53:00.000-05:002006-11-17T16:53:00.000-05:00I think you have this right. The result for the P...I think you have this right. The result for the Presidential election in 2004 was even more dramatic. I doubt the data is still available but I will recount this from memory. As exit polls came out showing a Kerry win the Bush contract went below 10%, then as each piece of news came out over the course of the evening the contract rallied to greater than 90% Bush win likelihood. But as you have observed, that market was following the news, not anticipating it or predicting it.Unknownhttps://www.blogger.com/profile/18246965714613827247noreply@blogger.com