Monday, December 12, 2011

Technical analysis doesn’t work

As a private trader, I’ve spent decades studying and using the tools of technical analysis in an attempt to improve my ability to anticipate the markets.

There have been times when my charts have looked like miniature Christmas trees set up by someone with an ornaments obsession.

10, 20-, 50- and 200-day exponential moving averages.

The parabolic sar and Person’s proprietary signal.

The MACD and the fast stochastic.

Fibonacci retracements, trend lines, price channels, the ichimoku cloud.

Not one of them is real. They’re all decorations that give the chart a look of professional wisdom and depth, but they are next to useless when it comes to understanding what will come next next.

Call me slow, but after all this time, I’ve finally concluded that the tools of technical analysis work just fine, except when they don’t. And when they stop working, it will be without advance notice, without a hint that something is about to go awry, and it will cost the trader money.

This is a radical statement, especially coming from someone like me who has lived and breathed the heady byzantine air of tech analysis.

So take your chart, and along with me, start removing technical tools. Strip them away, until you’re left with a chart showing each day’s open, high, low and close, and the volume.

It looks much cleaner, doesn’t it. Like a zen garden in Kyoto.

I’ve come to this conclusion through years of observation. Anyone who has read my analysis over the past two years knows that I’ve often used the word “whipsaw” -- 263 times. “Whipsaw” is just another way of saying that the technical indicator said one thing and the price said another. It is a synonym for “Oops!”

See enough whipsaws, and you can’t help but wonder what’s really going on. Since the market is, by definition, always right, then it must be the indicator that’s wrong.

Perhaps the biggest eye-opener was the Turtle Trading method, which uses Donchian channels -- the highest high and lowest low for the past 55 trading days.

Breakouts beyond the Donchian channel was said to show that there’s enough momentum to indicate a good push to the upside or the downside. But more often than not, I found that the Donchian channel was in fact a reversal point.

If a breakout brings with it the possibility of both momentum and reversal, it’s obviously a useless tool. If a technical analysis regime cannot be correct at least 90% of the time, then it a mere distraction.

The biggest problem with unreliable tools is that they paralyze the trader. “I see the signal, but is it a whipsaw? Maybe I should wait one more day...”

Imagine if you stepped on the gas in your car, and half the time the car moved forward, and the other if moved back, with no way to tell which it would be. I suspect you would get rid of that car in no time at all.

It is easy to conclude that indicators don’t work, but harder to say why. My theory is that most indicators are based on aggregating the past -- it’s all they have to work with -- and so they are continually out of date. New influences come into the market and the price turns on a dime; the indicator, working with averages and cumulative sums, takes awhile to catch up.

Having reached this point, I’m standing in the smoking ruins of my trading assumptions, grasping my last support, my lifeline: A simple price and volume chart.

My challenge is to figure out the best way to use this bare chart in trading, without the artificial technical analysis overlays.

1 comment:

  1. I tend to think TA is best for defense, and fundamental analysis is best for long term offense. TA is not worthless, it gives you five gifts. It measures what the market has been doing, spots divergences, makes predictions, structures a trade in the form of an entry point, profit target, and stop loss, and finally, TA tells you when you are wrong. Of all these gifts, knowing when you are wrong is the greatest, and fundamental analysis can't do this very quickly, if at all.

    TA just isn't very predictive. So what? We still need it for what it does.