Well, when technical analysis started to be disseminated, computers were not available to test the patterns, theories and indicators. It was a visual art. Remember that old book from John Murphy, The Visual Investor?
There were no backtesting results to prove or disprove the claims,
"Technical analysis methods of the old school include mainly chart patterns and some hypotheses of market behavior, for example Dow Theory. We have to keep in mind that personal computers were not available when technical analysis was first conceived and traders could not perform quantitative evaluation of claims made."
"(...) classical technical analysis was a “visual” method based on charts."
"During the early years of technical analysis, when prices moved below the neckline of a head and shoulders, for example, due to high autocorrelation in returns, the probability of a continuation of a trend reversal was high and the actual pattern was irrelevant at that point."
Is there anything really to a Head & Shoulders pattern? I do not think so. Peter Brandt might disagree, but I do not think it is possible to create a profitable trading strategy around these ancestral techniques.
(excerpts are from the article Why Some Technical Analysis May No Longer Be Effective (Forbes))
PS: Trading Course is now available both in Paperback and Kindle eBook.
