I just attended a one day seminar in Chicago covering the topic of prediction markets. So what are they? PM's are securities based on the possible outcomes of an event or question. The wisdom of crowds, or crowdsourcing is used to try and discover the probability that the event in question will occur. This has been used in sports betting for as long as anyone can remember, but is now being used to predict all sorts of things. The popularity of this phenomena has risen exponentially since the publication of James Surowiecki's book, The Wisdom of Crowds. Another good, although unstructured, source of information is Chris Masse's site.
I've logged a few of the notes here that came out of the meeting:
- Market participants can bet on future outcome of question or event, question is packaged as a stock in a market, betting is buying and selling the stock.
- Learns by weight updating like a neural net: those who are right get rewarded (wealth), and those who are wrong get punished, so on the next round, those with good info have a higher weight (more wealth to put in “stock”)
- Reasonably good predictors when majority have better than even chance of getting it right. Even if this is not the case, those with low chance get weeded out and weighed down in early runs, and those with good chance get wealthy, until good players control the wealth-weighted majority of the vote
- Important to provide enough incentive for participants to try their best to predict outcome
- There has been limited success with some internal trials by companies (Corning, HP). Other examples are the Iowa Electronic Markets, InTrade (run by TradeSports), and the Hollywood Stock Exchange.
- The market requires enough traders to provide liquidity in order to work. How many is enough. Someone's study showed that it didn't have to be a large number, just an active and informed group, perhaps as small as a dozen?
- Don't charge participants to play. If you do it will be subject to gambling laws.
- Rewards + Recognition + Relevance = Participation
6 comments:
Thanks for your review of the Chicago event.
How would I better "structure" my website?
Ah, I recognize that scene. In fact, I have a photo that looks almost exactly like your's. I'm guessing the seminar was at the Sheraton, right?
Scott, the seminar was actually in the Gleacher Executive Center at the Univ of Chicago.
Chris, it's a great and comprehensive site. It just seems to be a bit of stream of consciousness. You could organize by the following categories: Definition, Benefits, Working Examples, Vendors, Papers
Tim, it's a rich idea. I will mind this suggestion. Thanks.
Thanks for the report. Can you clarify or provide more detail on this point: "Don't charge participants to play. If you do it will be subject to gambling laws."
So why exactly isn't Everquest subject to gambling laws? They charge participation fees and players can win varying amounts of play money -- which is actually a "thing of value" in secondary markets.
Presumably it's because the real losses to players are capped by the participation fee. So why would charging a participation fee subject you to gambling laws in itself as your wording implies?
Hi Tim,
You criticized my classification at my prediction market directory.
I have a group blog up, on prediction markets. Take a look and tell me what you think of the categories.
I can take any criticism.
Thanks,
Soon,
Post a Comment