Does Technical Analysis Work?

Technical analysis is a very popular tool for many traders who use it to try to identify trading opportunities and patterns to profit from them. In a nutshell technica traders attempt to figure out which direction the price of a market is heading in the forseeable future, in particular whether the price is following an existing trend or a reversal or range trading. They do this by looking at past performance to predict future price movements, and take trades based on the extrapolation in price activity.

I have no problem with technical analysis skepticism, particularly with the more esoteric technical analysis like Fibonacci or Elliott Waves, it’s the holier than thou attitude I take offense to. As it happens, my heroes, in no particular order, are Isaac Newton, Richard Feinman, Albert Einstein, IK Brunel and Winston Churchill, so I’m a big fan of the ‘scientific method’…

Let’s not go over old ground again. A scientist would say:

a. believing something to be true does not make it true
b. If a theory predicts an event correctly on one occasion, it does not mean it will continue to do so
c. to test if theory is a useful predictive tool, you have to to use it to make a precise prediction which you then test by experiment.

So to convince us that chartism and technical analysis works you would have to carry out an experiment in which e.g. 100 chartists would be given 10 anonymous charts and asked to predict a share price in 3, 6 and 12 months for stating 95% confidence intervals for each value. These results should then be compared with the predictions of 100 non-chartists working blindly with the same anonymous graphs. The results should then be peer reviewed and published.

Technical analysts assume that the price of an asset is a reflection of everything that might be important to it, and is included in the present price. The problem with many chartists (not digging at you here) is that the predictions they make are so imprecise and qualified with so many caveats that they can always, in retrospect, be shown to be true. It is easy to make predictions that are always true (either side could win, or it might be a draw) but such predictions are valueless.

e.g from the IC a month ago, made when cotton was about 180:

“We have Fibonacci-projected target for cotton at 261. However a correction has started at 209.97, which could see cotton unravel (sic!) to below its lower blue channel line and towards 168.72. That represents the 38.2% Fibonacci retracement of its major up move to date, where it may find support. However, a break of that level could see it enter an even more prolonged pull-back phase.”

Which in jargon-free English says: It could go up from 180 to 261, but it could go down to 168, from where it could stay or go lower. A prediction that will always be true. It actually went down to 150 or so in a short time, a major move to 60% or so below its “Fibonacci target”.

So it’s no use trying to persuade a true scientist that technical analysis works by quoting your own experience and beliefs. It’s no good.

I’ll shut up now.


I think many who predict major trend changes do it for enjoyment because of the kudos of calling it right. I don’t think as many actually base a long term position off the back of a prediction based on technical analysis until some form of confirmation occurs. (admittedly I have done this, but the stakes were small).

Having said that technical or fundamental analysis, or a combination of the two, remain vital if anyone wishes to make a consistent profit in this fast moving, challenging market.

  1. No comments yet.
  1. No trackbacks yet.