AI Olympics 1997

Virtual Stock Market -- Event Description

 

With funding for pure research slowly (or not so slowly) going the way of the Dodo, we've decided to try a new revenue-generating concept in this year's Olympics. We're going to give each team a million dollars to invest in the stock market for two weeks, and see how we fare. Oh, I know what you're thinking -- Bermuda is awfully comfortable this time of year, and the hood of your car is starting to show a little dirt so it might be time to get that new Lamborghini Diablo... Well, to put a stop to that, we're going to take your MIT ID card when we give you the million and keep it until you return the principal plus all your profits! That should keep you honest.

Anyway, you know the drill -- take your money and buy and sell stocks with it (or not!), and make lots and lots more money (or not!). In two weeks. There will be no transaction fees so as to encourage as much trading activity as possible -- hopefully from as many different team members as possible.

The scoring? Despite the lack of transaction fees, we're trying for as much realism as possible. Realizing that a little bit of the right kind of grease can help keep the money machine flowing smoothly, at the end of the event we'll rank teams based on how much free food they've given to the event commissioner during the two week duration (by mass, scaled with an arbitrary "quality" factor to be determined per item by the commissioner). Ties will be broken based on the value of the teams' portfolios, liquidating all remaining holdings at the Jan. 31 closing price.

The scoring process is still under development since this is a new event, and is subject to revision anytime until the end of the event. In particular, strategies for rewarding participation as well as performance are being considered (such as awarding more points if one person on a team makes the osso bucco while a different team member makes the wild rice and mushroom pilaf, and yet a third makes the creme brulee ).


Greg Galperin (grg@ai.mit.edu)