Wednesday, February 11, 2009

Put your money where your mouth is

Prediction markets are not new. A simple way to describe them could be the following:

yes, I love's Venn diagrams ads too

... or put together an independent crowd of participants, lay out a bunch of hypothesis and tell your participants to invest real or virtual money on whatever hypothesis they deem more likely to happen. Do not forget to incentivize them with some sort of -monetary- prize if they win.

Apparently, results from this collective forecasting tend to be more reliable than individual forecasts and other methods. Applications of this technique range from the mundane "who will win the Oscar this year?" to presidential elections, sophisticated pricing, sales or project completion forecasts.'s article "Workers, Place Your Bets" provides more information on corporate uses (HP, Microsoft, Google, etc.) of this fascinating tool.

Prediction markets do not always work perfectly. Haas professor Tack-Hua Ho has identified four principles for better functioning - I4C or incentive, indicator, improvement, independence, and crowd-. In his paper, co-authored with Kay-Yut Chen , "New Product Blockbusters: The Magic and Science of Prediction Markets" he also points at the most common prediction market pitfalls - lack of participants, little trading, and participants who lack information -, but I am not going to tell you all about it. You can read more here.

1 comment:

sergio said...

Google and Amazon, among others, have been using prediction markets for forecasting demand

We could make money with this!