New Prediction Site: Predictify.com

Predictify.com launched in early October as an attempt to provide a simpler and more open version of prediction markets such as Intrade.

There are several advantages I see to their approach, primarily around its relative ease of use for end-users.  More traditional prediction markets that are based on buying and selling of shares bring with them the inherent complexities of such a system.  Predictify, instead, follows the “guess the number of jellybeans in a jar” model where each person simply guesses that they believe the correct answer is and are rewarded based on accurate their answer is.  For a public-facing tool, I believe the ease of use will keep the user base large and therefore produce more accurate answers.

What I like best, however, is that they support both free and “premium” questions.  Free questions are limited to 100 responses from the public, but are a great way for companies to experiment with the tool and its accuracy.  Their incentive system includes an incentive for users to answer free questions which improves their accuracy.  With the premium questions, the number of responses can reach 10,000 and includes profile information on the predictors.  Furthermore, the fact that money can be made from answering premium questions correctly causes the quality of the predictions to rise, as well. 

One example of a current premium question is to predict the sales of Scott Adam’s upcoming Dilbert book.  For the question price of $5,000, the book’s publisher is gathering extremely valuable sales prediction data which should prove to be very accurate.