Technical analysis doesnt work

Doug Reed, the CTO and Founder of Cake Financial, recently posted the question in his network: Does technical analysis really work? It doesn't, at least as applied by home traders because:

1. Technical analysis assumes that past prices can sometimes predict future prices. But the prediction is made by LOOKING AT A CHART. The market certainly has a few irrational people that lead to some small, temporary inefficiencies in price. But there are also gigantic funds using automated trading to trade out exactly these inefficiencies. There's no way that looking at a chart can find inefficiencies as effectively as a bunch of statisticians with a Cray.

2. The day to day fluctuations in prices are just noise. The only sustainable up or down trends are from fundamental economics - companies that make more money over time are worth more over time than companies that don't.

3. Technicians who have been successful using their method are no proof of the method's success. If you traded Google (GOOG) over the last few years and are holding it now, you made money regardless of why you traded it - whether because of fundamental analysis or technical analysis or astrology.

There are two flaws with looking at successful technicians as proof of the methods: (1) there's a major survivorship bias. People who have lost money for long periods using technical analysis no longer use it. So you don't see the losers of the technical analysis game because they've exited. You'd need to include them for the data to be worthwhile, and (2) correlation does not equal causality. The statement "technical analysis doesn't work" is not the same as "you will always lose money using technical analysis". If saying technical analysis didn't work were the same as saying it always lost money, the answer would be easy - just do the opporsite and always make money. You may make money using it and you may lose money, but the technical analysis won't be the cause of it. Lots of people have systems for roulette, even though there's no mathematical way to gain an advantage at a fair wheel. Some of them make money anyway. That doesn't mean their system works, just that they were in the right place at the right time.

4. Any strategy with frequent trading is going to lose so much to the friction of taxes and fees that it will be especially difficult for it to outperform substantially.

Thanks to Nassim Taleb in Fooled by Randomness for some of the examples.

(As an aside, I still really like Cake, even more after meeting the team right before my vacation. But please, PLEASE make it easier for the investing discussion to carry on in the existing online communities - blogs, Facebook, etc. There's no reason I'd want to post a long response to a question in the Cake walled garden and then repost elsewhere... A simple widget that lets me ping users in my network with financial content I find interesting combined with a mechanism to send my questions directly to the network (instead of posting them) would make the discussions much more vibrant.)

Technorati: technical analysis, investing, cake financial