Long & Wrong

If you google “head and shoulders pattern” then it returns 543,000 pages, only a few of which are about the shampoo.  One of the sites on the first page of results describes it as a “powerful reversal pattern“, but is it?

You have to dig very deep to find any kind of quantitative answer to that. This strikes me as strange, there is no shortage of data on which to test the idea(!) and no difficulty formalising the pattern as an algorithm.  Why is it so difficult to find an analysis of the value of this pattern (and indeed many patterns and indicators)? The cynic might answer “because it doesn’t sell books“.

But I did dig deeper and found this study, Head and Shoulders: Not Just a Flaky Pattern (you can download the pdf of the paper), by Carol Osler.   I’d really be pushing your attention span, and mine, to critique or adequately summarise this paper, you can however read a journalist’s slightly uncharitable interpretation here.

Head & Shoulders

My point is not about this paper in particular but rather the general observation that quantitative evaluation of technical analysis is noticeable only by its absence, or at the very least its scarcity.  It only takes a cheap PC and a free download of OpenOffice Calc to be able to do a rough and ready back-test  for yourself.

During a heated debate on a trading forum about the value of pin bars I did exactly that. I analysed OHLC data for 8 years of hourly EUR/USD data and found they had no (actually slightly worse than ‘no’!) predictive value.  The problem of course is that crunching numbers in this way doesn’t allow for these signals to be placed in context, that limitation doesn’t however entirely negate the result. What was most telling was when people disproved  my analysis by showing me an example when a pin bar had bagged them 70 pips(!).

So my hope is that we become more prepared to shine the light of analysis on our cherished beliefs, and equally prepared to ditch any that don’t stand up to the scrutiny.

I have provided a number of links in the text above where I feel some terms may need further explanation.  I know the vocal minority are smart enough not to need them, please  don’t feel patronised.

The steadily growing traffic for this blog (thanks to those that kindly link here) is only a sign that new people are visiting, they may be visiting and hating it, if so I’d like to know. The silent majority are encouraged to comment, as of course are the vocal minority, even Solfest.