Monday, September 27, 2010

Lo and Hasanhodzic, The Evolution of Technical Analysis

Long before I had ever heard of technical analysis I was involved in the production of a book that I considered the most obscure in a field of pretty obscure books. The exact title escapes me, but it was about Babylonian accounting. Little did I think that I’d ever revisit Babylonian financial recordkeeping. Then came the new book by Andrew W. Lo and Jasmina Hasanhodzic, The Evolution of Technical Analysis: Financial Prediction from Babylonian Tablets to Bloomberg Terminals (Bloomberg Press, 2010) where I learned that over the course of four centuries Babylonian diaries recorded the market prices of the same six commodities—barley, dates, mustard/cuscuta, cress/cardamom, sesame, and wool. Fascinating! It all goes to show how my own interests have evolved.

The Evolution of Technical Analysis takes the reader on a whirlwind trip through history and across continents. Starting in the ancient world, progressing through the Middle Ages and the Renaissance, moving to Asia and then to the new world, the book proceeds to the “new age for technical analysis” exemplified by such luminaries as Dow and Gann, and finally looks at technical analysis today. It then offers a brief account of randomness and efficient markets and summarizes some academic approaches to technical analysis.

The history is largely reliant on secondary sources, but the authors have combed through the material to trace links between technical analysis and astrology,* highlight parallels between the writings of the eighteenth-century Japanese rice trader Munehisa Homma and Charles Dow, explore people’s penchant for manipulating markets, and showcase principles of technical analysis that have withstood the test of time. Their conclusion? “Technical analysis has had to undergo so little change precisely because it is so robust and so deeply relevant to how markets operate.” (p. 105)

Nonetheless, markets have evolved and continue to evolve. Indeed, “during the last three decades the market has been evolving so rapidly that the very intuition on which developing technical or quantitative trading strategies is based—the stronger the historical backtest results, the more promising the strategy—got turned on its head. The rate of change has been such that historical viability has become hardly indicative of present success, challenging today’s portfolio managers to develop adaptive strategies capable of detecting shifts in the market environment.” (p. 119)

The authors contend that pattern recognition may be the key to dealing with changing environments. The results of an online study that I referenced earlier on this blog “provided overwhelming statistical evidence (less than 1 percent probability of the findings being generated by pure chance) that humans can quickly learn to distinguish actual price series from randomly generated ones.” (p. 146) Since human pattern recognition skills outstrip those of any known algorithm, it may be possible, they argue, with the proper interface to “translate the hand-eye coordination of highly skilled video-gamers to completely unrelated pattern-recognition and prediction problems such as weather forecasting or financial trading.” (p. 129) I wonder what industry-standard terminals will look like in 2050!

*The authors find “nothing in favor of astrology’s inherent predictive value.” But, they continue, “For our predecessors of precomputer eras it may have seemed justifiable to use astrology as an input in their price-based forecasting models. In fact, astrological inputs may have been genuinely useful to prediction insofar as they played the same role that random-number generators play in today’s forecasting models.” (p. 104)


  1. Very helpful thank you

  2. Dear Brenda
    I decided to purchase the book after your favourable review.
    Although Lo and Hasanhodzic present a good historic account of finance and speculation, I feel that their attempts to link technical analysis with some of the very shaky evidence from the ancient times very far-fetched. They really did stretch their evidence to its limit in order to link it up to their hypothesis.
    Although interesting I feel that the prehistoric sections could have been reduced to a single chapter.
    Would be interesting to hear your views.