Commentary

Robert Schuessler
Traders Magazine Online News

A Smarter Monkey

In this contributed piece, TIM noted that some traders do better than others when using data that has been run through certain analysis - that is, have used some form of machine learning to assist them.

Traders Poll

In his first public speech, SEC Chair Jay Clayton deviated from his prepared remarks and offered his own "off the cuff" comments on market issues. Do you like this change of pace?




Free Site Registration

August 31, 1998

The Smartest Way to Learn Trading? Win or Lose:Trading in the: Trenches With I-SMARTS

By John A. Byrne

Also in this article

  • The Smartest Way to Learn Trading? Win or Lose:Trading in the: Trenches With I-SMARTS

Staring at his computer screen, a trader watches his position heading south. Stock XYZ is tanking. About ten seconds later, a second dealer offers 1,000 shares in XYZ, up 1/16. The trend is reversing.

The stock's underlying value slowly moves north, driven by informed traders who react to market fundamentals, such as positive-earnings surprises.

The first trader then unloads his position, making a modest profit. Cool as a breeze, he allows himself a sharp guffaw before plunging back into the market.

What the heck, he tells himself, buy another 1,000 shares of XYZ. Now, the market makes him modestly rich on paper. Boy, was that easy.

A trader in another environment would have given high fives to his buddies. But not this guy. The problem or perhaps the beauty is that he's playing a game on I-SMARTS, the interactive securities-market trading simulation.

The trading system is the brainchild of Robert Schwartz, finance professor at New York's Baruch College, and Bruce Weber, assistant finance professor at New York University in downtown Manhattan.

I-SMARTS, which is DOS-based, features five fictitious traders each with funny names, such as Tuna and Dog. And the buzz is that it is the only system available that adequately replicates several separate markets: quote-driven markets for dealer orders; order-driven markets for customer limit orders; and a call auction with periodic matches.

I-SMARTS is not perfect, of course. It is not networked and is only programmed to trade a single stock, which has no fundamental information (though it does post the prices on its last 15 trades).

The system itself really only allows trading in 99 units, or 9,900 shares, though the live trader could enter a block order of more than 10,000 shares. But that would trigger market impact and render ongoing trading meaningless under current parameters.

The absence of a SelectNet and SOES functionality for dealer trading, as well as orders residing on simulated electronic communications networks, might also be viewed as a limiting feature.

But the game tries hard to be realistic. Each trader starts with no cash, and can accumulate long or short positions over several hours. A market index gives traders a performance measurement. (A history of a trader's decisions is available).

To stimulate traders' attention under research conditions, Schwartz and Weber give out cash authentic legal tender based on their performance.

"After half an hour of trading on I-SMARTS, I said to one guy from the industry, You've made enough. We'll buy a Big Mac with your $5,'" Schwartz recalled with a laugh. "My experience is that it is not easy to make money trading."

The two academics developed simulations that are not purely random, meaning that buying and selling pressures, for example, do not strictly follow textbook economics models, but patterns analogous to real-world conditions.

Experimental Conditions

While the academic field of experimental economics has cranked out volumes based on studies of networked or electronically-connected participants playing the roles of traders entering bids and offers, I-SMARTS is distinctly different.

Three marketplaces can be commingled simply by clicking on a menu. Equally important, it is programmed with order-flow scenarios and other market parameters that would test the nerves of the most seasoned buy-side and sell-side traders.