Friday, December 8, 2006

Humans, Chess Computers and Black Box Traders

On December 5, 2006 it was announced that a computer, Deep Fritz, beat the world champion human chess player Vladimir Kramnik, in a six game tournament. Four of the games were draws and two were won by Deep Fritz.

It is also believed that over half (estimates vary) of the trading volume in stock markets is now made using machines, often called black boxes. And let me be clear: these are machines that are programmed to evaluate stocks and make trades without human intervention.

Given that, would individuals be best off turning over their investments to mutual fund managers or hedge funds that can make use of these AI (artificial intelligence) computers?

Well, some individuals clearly were not up the the task of competing in the markets even before machine trading started. For the rest of us, so far at least, I think that if you take a look at the nature of machine trading, you can avoid being hurt by it and might even take advantage of it, at times.

What are the machines doing? Their strategies mainly fall into four categories: momentum plays, arbitrage, covariance, and news analysis. All of these strategies were developed by humans, but computers can be better at them.

Momentum plays are the simplest strategy: if a stock is going up, bet that it will continue to go up. If it is going down, bet that it will continue to go down. Be ready to bail out of a position the moment momentum changes. This simple reality used to be the main advantage of having a "seat" in a trading pit. What are the advantages for a computer here? At best humans work in tenths of seconds. Computers can work in millionths of a second; in a second a computer to look at the entire range of stocks to see which ones make the best momentum plays any given moment.

The result of course, is amplification. Some people say stocks are always in equilibrium, the latest price representing a balance between the buyers and the sellers. But it is just as true to say that stock prices are always out of equilibrium. If a stock is going up it attracts momentum players, human and computer alike. It goes up until it is so out of touch with fundamentals that other computer programs and humans start selling it in quantities sufficient to flatten the curve and then send it heading down. The same process in reverse runs on the downside.

Are you thinking about buying a particular stock today? You know if it is falling you are inclined to wait and see if you can get it cheaper. If it is going up, you buy quickly because time is money.

You don't have to worry about the machines too much if you are an intermediate to long-term investor. Ultimately stock prices depend on fundamentals. While some machines are throwing prices out of equilibrium, others are executing trades that tend to bring prices back to fundamentally sound levels.

So do your research, delve into the future, and try to buy at a good price.

I'll be looking at the other three strategies (arbitrage, covariance, and news analysis) in future blogs.

No comments:

Post a Comment