Mercury’s Blog

If you're new here, you may want to get Mercury's Blog by Email or subscribe to my RSS feed. Thanks for visiting!

Initial conditions are critical in the design of a new prediction market. The potential trading that can be done by traders in their first few experiences with the site can foster behaviours that encourage them to continually check back and trade further, or can contribute to behaviour where they never visit the prediction market again. There are a couple things to consider in the design.

Initial Stake

How much money will the traders have when they first log in? (This assumes that traders use play-money or are given equal amounts of real money to use.) There are different ways of approaching this. Inkling Markets, for instance, gives each trader $5000. Assuming that each stock is at a mid-point price of $50, they can only afford approximately 100 shares with their initial stake. The default options for purchasing shares at Inkling are 5, 20, and 50 shares, meaning that it is entirely possible that if a trader feels very strongly about some particular contracts, they will use up their entire stake trading in just two of them.

The Hollywood Stock Exchange has gone to the other side of the spectrum. They stake each trader with $2,000,000. A standard MovieStock is priced between $20 and $200, with default purchase options of 1000 contracts or 25000 contracts. With this initial stake, a new trader could purchase 20 different MovieStocks, all priced at $100 each, at the default lot price of 1000 contracts. That is a LOT of trading, and it is very likely that a new trader won’t use it that much in the first instance, thus causing them to come back to the site and trade again.

Timescales of contracts

The problem with some prediction markets is that most or all of the contracts on the site won’t be judged until some time well into the future. This provides little incentive to keep coming back to the site, and drives liquidity down. In these cases, it can be very valuable just to have a small set of contracts that expire in the short-term, even if they’re unrelated to the main objective of the site. For instance, MediaPredict is trying to find the hot new books and bands, but the contracts for the book contest wouldn’t be judged for several months after the site went live. Since then, they’ve offered contracts on events such as “Will Interpol debut in the Billboard Top 5,” “First day box office total for Harry Potter,” etc. This drives continued turnover and interest in the site, keeping the long-term contracts visible and more liquid than they otherwise would have been.

Again, the Hollywood Stock Exchange is a great example here as a trader can participate at any stage of a film’s creation, from pre-production all the way through to the first days and weeks at the box office. Since new films are out every week, there are always long-term and short-term contracts on which to speculate.

What this means for you

As always, your choices here are tied to your goals of your prediction market. Inkling is likely more concerned about manipulation, particularly in small user-generated markets, and thus limits the effect that any brand-new trader can have. The Hollywood Stock Exchange is big enough that no new trader is going to have much of an effect, and they are far more concerned about getting people to trade in a variety of contracts (thus providing more data that they can sell to the film studios).

Your choices will be affected by the number of people participating, if they are specifically selected and knowledgeable (or are general employees that join), and the information you’re deriving from the prediction market. Your choices for initial conditions are tied up in the overall goals for the project. As always, implementing a prediction market is a bit of an art as well as a science.

———-

Please contact Mercury Research and Consulting to discuss how we can help you structure your prediction market project.

This entry was posted on Tuesday, July 10th, 2007 at 10:25 am and is filed under General.