Saturday, September 10, 2005

Risk management


  • "I would say that risk management is the most important thing to be well understood." -- Bruce Kovner, page 82, Market Wizards by Jack Schwager

  • Jack Schwager: "How much of a role does luck play in trading?"
    Richard Dennis: "In the long run, zero. Absolutely zero. I don't think anybody winds up making money in this business because they started out lucky."
    Jack Schwager: "But on individual trades, obviously, it makes a difference?
    Richard Dennis: "That is where the confusion lies. On any individual trade it is almost all luck. It is just a matter of statistics. If you take something that has a 53% chance of working each time, over the long run there is a 100% chance of it working. If I review the results of two different traders, looking at anything less than one year doesn't make any sense. It might be a couple of years before you can determine if one is better than the other."

    --Page 97, Market Wizards by Jack Schwager

  • "Don't focus on making money; focus on protecting what you have." -- Paul Tudor Jones, page 139, Market Wizards by Jack Schwager



A couple weeks ago I was reading the "New Era thinking" chapter in Robert J. Shiller's Irrational Exhuberance, the original edition that was published shortly before the nasdaq crash of 2000. Shiller found that media buzz about an emerging "New Era" historically happened before and during large stock market advances that ended with crashes. So naturally I bought a copy of Visions by Michio Kaku when I saw it in the bookstore. The significance of Visions was its publication date, 1997. Kaku wrote Visions as a guide to the cutting edge of science at the time, and as a roadmap of how we could expect our lives to be changed by emerging discoveries.

Kaku is a physicist himself. The book jumped out at me because I remembered his name from a mind-blowing article Kaku wrote and published on Ray Kurzweil's site that I had read several years ago. One of the most interesting passages I've seen so far in Visions was this one:

  • "If these difficulties in computer technology can be overcome, then the period 2020 to 2050 may mark the entrance into the marketplace of an entirely new kind of technology: true robot automatons that have common sense, can understand human language, can recognize and manipulate objects in their environment, and can learn from their mistakes." -- Michio Kaku, page 16, Visions


What I find to be so interesting about this statement is that there may not be sufficient demand for these automaton housekeepers. The fact is, it is so much easier to create software "risk managers" of the kind talked about in the above quotes from Market Wizards than it is to create robotic servants.

As computer hardware and software technology continues to advance, the financial market computer models and risk management computer software gets more sophisticated and more difficult for humans to outperform. The people maintaining and controlling the best of these software programs will control the world in a sense. The primary function of financial markets is to optimize resource allocation. These software programs are the ultimate resource allocators.

There will always be profit opportunities in the financial markets for humans, but let me put forth a scary scenario. Consider that software systems are responsible for more and more of the bidding and offering that goes on in markets, and at some point they might reach the point where they can have markets moving so erratically that traditional technical analysis loses all predictive value. Many readers familiar with the markets probably think it is absurd that this could happen, but let me introduce a science experiment called the Prisoners' Dilemma. In 1984 a tournament was held where anybody could write and enter a computer program to compete in games of iterated prisoners' dilemma. The winner of the tournament was a strategy called Tit-for-Tat. Tit-for-Tat was the dominant strategy for a long time, but more recently a program was written that could detect whether it was playing against a Tit-for-Tat program by analyzing its opponents' reactions to its moves. If it detected a Tit-for-Tat program, it would make moves that destroyed Tit-for-Tat's score, at a small loss for itself.

If you have sophisticated enough software controlling enough money in the markets, by collusion they could keep humans out. For example, if you don't flash the right series of bids and offers at the exactly correct time intervals, they could determine that your bid is a human bid. Or if your stop isn't rotating through the correct price series, then it must be a human stop, or foreign software and is now a target.

This is just another reason why money today is worth a lot more than most people fathom, because it may not be as easy to come by in the future. There are many patterns in the stock market today that provide profit opportunities and haven't been arbitraged out because it is too expensive to fade them. One makes more money by amplifying the pattern than disrupting it, but the risk vs. reward ratio for any given pattern is certainly subject to change over time. Higher failure rates tip the scales in favor of the very best risk managers, whether they be human or machine.

2 Comments:

At 7:59 AM, Anonymous Anonymous said...

Superb post. I feel bad for all of the people who skipped over this one because it's kind of long.

 
At 8:10 AM, Blogger jontait said...

Thanks so much =)

 

Post a Comment

<< Home