Saturday, June 14, 2014

Making Technical Analysis Dynamic

A while ago, I posted on the topic of divergences appearing in the U.S. stock market.  Specifically, we were near all-time highs in the large cap averages, but a significant number of stocks were making new lows and small caps in particular were dramatically underperforming.  When I examined past instances of such underperformance, I found very different results depending upon the VIX regime that we were in.  When divergences occurred in high volatility environments, the forward results for SPX were bearish and volatile.  When the same divergences occurred in a low VIX environment, the forward results for SPX were actually nicely bullish.

The same results could have been found if I had been looking at chart patterns, oscillator readings, or other trading "setups".  It's not that they lack value; it's that their value is contingent upon the context in which they occur.  Should you buy after a couple of days of strength?  In a low volatility, range bound market, the answer might be quite different than in a rising volatility market displaying momentum characteristics.

This is a limitation of how traders tend to implement technical analysis.  Too often we assume a static reality, so that a given chart pattern, oscillator reading, or Fib level has a fixed meaning and significance.  Psychologists tend to be skeptical of static depictions of reality.  Most human interaction is context-dependent:  someone reaching out to hold my hand crossing a street means something different than the same gesture in a hospital room.  Or, as the old joke has it, "bear right" means one thing in a car, another thing on a hunting trip.

Challenges in anticipating market movement may be a function of our need to find fixed setups.  Looking for the same patterns in very different markets might be a formula for temporary trading success.

I strongly suspect this is an important topic.  

Connie Brown was on the right track when she proposed that oscillators behave differently in bull and bear markets, requiring different interpretation.  As many technicians have noted, oscillators themselves are more useful in certain market conditions (range bound) than others (strongly trending).  John Ehlers and Ric Way compute dynamic cycles for stocks and indexes, with frequencies and amplitudes that wax and wane with shifting market conditions.  Their use of quantification to turn static indicators into dynamic ones also strikes me as quite promising.

But what if, as market regimes change, fresh technical indicators gain predictive value and others become less relevant to forward price movement?  In such a dynamic world, "setups" would always be evolving; you'd always be learning markets.  Your edge wouldn't be a core set of trading patterns, but your ability to identify and trade the patterns that possess an edge here and now.

The implications for coaching are significant:  Advising traders to "stick to your plans" and "follow your process" works as long as market regimes are stable.  Then it doesn't work.  In a world of changing markets, adaptability is the new discipline.

The implications for mentorship are also significant.  Teaching the same chart patterns and technical rules at all times to all traders is like using the same training for soldiers who will be performing in the desert, at sea, and in rainforests against established military forces at some times and insurgent guerrillas at others. If there's one thing elite fighting forces and well trained athletic teams know, it's that you study the opponent and adapt your strategy and tactics to the situation.

How many traders truly study markets and adapt their strategies and tactics to the threats and opportunities they identify?

In the next post, let's take a look at what a dynamic technical analysis might look like.

Further Reading:  What Are You Doing Between Trades?