Tuesday, February 06, 2007

Deepening Our Understanding of the Stock Market's Personality

When we refer to the "personality" of an individual, we're creating a shorthand that describes that person's tendencies: How they think, feel, behave, and relate to others. Our categories of personality--such as calling someone "Type A" or "narcissistic"--help us understand that person, predict their behaviors, and communicate our understanding to others. To be sure, personality traits don't capture all of the variance in human behavior; much is determined situationally. Still, human behavior is sufficiently patterned that we benefit from our various categories.

When we perform studies of historical market behavior--or when we create a mechanical trading system--we are attempting to capture some aspects of the stock market's personality. Each study or system is looking for a recurring pattern. We might say that the sum of patterns across multiple time frames defines the personality of a given market. Some markets have "trending" personalities; others are "volatile". With markets, as with people, such personalities don't capture all the variance in the behavior of a stock or index. Situational factors, from interest rate movements to geopolitical developments, play important roles. Still, we see sufficient patterning of markets that it is meaningful to say that, for example, we've been in a low volatility bull market.

When we deal with a person, we do not expect that his or her personality will radically shift from day to day. The tendencies that have stood out in the recent past are assumed to hold for the immediate future unless we have some reason to believe that the person has changed or that the situational influences are unique. Similarly, traders assume that past patterns will carry forward to the immediate future. This is just as true for discretionary traders as mechanical systems ones: whether we trade off of chart patterns or quantitative relationships, we are assuming that patterns we've noticed in the past will, in the absence of situational influences to the contrary, carry forward to the near-term future. Without that assumption--the assumption of some patterned nature to markets--trading makes no sense whatsoever.

We tend to think of discretionary trading--trading based on the real-time judgment of the trader--and mechanical trading--trading based on a defined system of entries, exits, and money management--as opposite sides of the trading spectrum. In reality, however, both are undertaking the same thing: capturing market patterns. It is for this reason that the development of trading systems greatly benefits from the expertise of traders who employ their judgment and discretionary traders benefit from the decision support of knowing quantitative market patterns. As I suggested in a post a while back, the ideal is a trader with both head and gut: someone who has a feel for market patterns and a head for market odds.

During my morning trading sessions, I have sometimes referred to the Odds Maker module of the Trade Ideas screening program as one bridge between discretionary and system trading. Indeed, today's Weblog entry mentions an Odds Maker pattern regarding downside breakouts of opening trading ranges. I am a discretionary trader, but I find it helpful to know that, during the past three weeks of trading, the great majority of downside breaks of the opening trading range have been reversed. Indeed, on balance, one would have made nice money buying such breaks rather than joining the sellers. Knowledge, indeed, is power.

What Odds Maker is doing is capturing a piece of recent market personality. A broader piece of market personality is captured by a trading system that has been tested over many years, such as the one recently mentioned in my post. Such patterns and systems won't account for all market behavior--their absolute edge might be relatively small--but followed consistently over time, they can yield major profits. The bullish trader who has bought market weakness has consistently outperformed the bullish trader who has bought strength: knowing the market's short-term personality has been crucial to the success of the active trader.

The bottom line? Two important implications:

1) Quantitative assessment can aid subjective judgment. The psychologist relies on psychological tests; the fighter pilot relies on radar readings; the poker champion knows the odds of each hand. Knowing patterns--whether of clients in therapy, opposing football teams, or markets--helps professionals make better decisions.

2) The expertise of subjective decision makers can aid the development of quantitative tools. I have seen many programmers at trading firms who have programming expertise and no feel whatsoever for *what* to program. It's when the expert trader shares patterns with the programmer that a superior trading system can be developed--the kind of system that, then, can offer the aforementioned decision support.

Can discretionary traders form partnerships with quantitative system developers for their mutual benefit? My next post will propose a vision for just such an arrangement.