Bluffing Could Be Common In Prediction Markets | Hacker News Reader

Bluffing Could Be Common In Prediction Markets

sciencedaily.com

This is an interesting but fairly esoteric point. It requires the following to be true

1. The bluffer knows some secret information that is not known by other participants.

2. All other market participants must believe the the bluffer is acting on secret information by their bet.

In a market with many users and anonymous trading I can't see it being worth the risk to bluff. There are bound to be traders out there who will simply accept your money when you bet for the 'impossible' result and any money you make by betting the other way is not likely to compensate.

With non-anonymous trading it is possible that a trader with a high reputation could pull this off but they would damage their reputation in the process.

Apparently, Nathan Rothschild used this gambit on the London Stock Exchange. He had a trusted confederate observe the battle of Waterloo, then rush to London the moment he determined the outcome of the battle. (Some hours before it ended.)

Nathan then received this news, then acted as if the opposite were true in order to manipulate the market to his advantage.

http://www.rumormillnews.com/cgi-bin/archive.cgi/noframes/re...

Interesting information. So it can work once. He obviously had a significant reputation before this and I'll bet people were a lot more suspicious of his moves in the future.

His reputation suffered in one respect, but he made so much money and gained so much power, it was worth it. In the eyes of many, his reputation as a shrewd and canny operator was confirmed.

I've done a lot of research/consulting in prediction markets. I'm not totally surprised that in theory bluffing is possible. I've seen cases of market manipulation in my research.

But I think everyone should look at the results in a different light: Are they correct??

If the results (either probabilistic results or index results) are accurate, then if there are minor cases of manipulation who cares? I don't believe that prediction markets are that different than any other marketplace with minor concerns of manipulation.

As for accuracy, I've written about that here:

http://blog.mercury-rac.com/2008/01/25/how-to-interpret-pred...

(most people, and virtually all journalists, just don't get the math when it comes to probabilities.)

PDF of the actual paper: http://www-personal.umich.edu/~rsami/papers/manip.pdf

Reply on news.ycombinator.com