Is it a good business proposition to run a prediction market?

My friend morrell pointed out these tweets from Michael Story of Good Judgment, Inc., a firm which sells forecasting services based on survey-format crowds. Michael tweeted:

Concidentally, about the same time I heard about this tweet, I got this message from

PredictIt had a very busy night last Monday as the results from Iowa were coming in. Not only did many of you want to place trades, but we had many other visitors to the site, who were watching your activities closely and refreshing pages regularly.

As a result, the site was often frustratingly slow to load. I want you to know that we are working around the clock to improve performance during these times of increased traffic. But it is going to take us some time to ramp up.

Granted, they are a funded nonprofit research project, but a fair portion of their funding comes from the sizeable 18% rake on cash-outs, and it looks like they are doing just fine.

Emile Servan-Schreiber, of Hypermind, had this to say:

I’m not sure what kind of prediction market he is talking about. Betfair has been doing very well since 1999 or so. Intrade failed only due to mismanagement, after 10 years of spectacular growth despite adverse US regulations. PredictIt is doing very well despite regulatory limits, and growing fast. IPredict in NZ had been doing fine for almost a decade until they had to shut down only because the gvt passed new regulations forbidding its existence. These are all real-money markets, which is a proven, robust business model.

Michael is right that public-facing play-money markets are much harder to sustain. When the participants don’t pay to play, the money has to come from somewhere else. Advertising is not a solution (not enough eyeballs), and sponsorships are hard to come by unless your predictions have a superb track-record. This is where the Superforecasters & Hypermind models (essentially the same) are promising. Instead of going after large crowds (like GJOpen), rely instead on a relatively small but carefully selected group of excellent forecasters and consolidate their forecasts for sponsors who really need them. Whether you use a prediction market (Hypermind) or a survey platform (GJ) to consolidate the forecasts makes no difference in terms of accuracy. There are pros and cons to each method, and individual forecasters may prefer one platform or the other depending on their availability, familiarity with markets, and appetite for competition.

So the question is whether a play-money prediction market which is a front for a market-research firm can do well.  My model of GJ, Inc. and its site GJOpen is that it is developing a McKinsey & Co-style market research business model.  They are leading with the hook that they are an IARPA child, but I don’t see them growing that way, in the same sense that Google behind the scenes may do some Government work (in particular they have a lot of natural language processing patents which are obviously of interest, plus the secure cloud business), but Google’s bread and butter is not the national security industry per se.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s