What about Technicals?
Market efficiency is often highly contested by proponents of technical analysis. At Ovtlyr, we consider technical analysis to be a rudimentary form of behavioral analysis. In a recent test, 74 (we stopped there because the results were so uniform, but we encourage you to try the following experiment for yourself) large cap stocks totaling over 2,300 combined years of public data – all constituents of the S&P 500 and most likely to demonstrate highly efficient pricing – were subjected to a MACD strategy as outlined by Fidelity. Across these assets in a long-only capacity, the average exposure (days holding / total market days) was just north of 60% while capturing 37.45% of returns relative to buying and holding over the entire duration, giving the strategy a 62.24% effectiveness compared to buy & hold.
The natural response among some at this point would be to decry the strategy as too simplistic, or applied to the wrong assets, “I only act on the MACD when it’s in alignment with [some other indicator] and not within 3 business days of a lunar eclipse…” perhaps these arguments are correct (we’ve personally seen some convincing cases), perhaps not, but the nature of the argument is fundamentally immaterial here. The purpose of using this methodology was due to its ubiquity. We specifically wanted a metric and assets seen by many market participants in order to effectively compare the same strategy against complete and blind randomness.
We constructed a random-walk generator, cumulatively adding or subtracting each successive day based on the distribution of outcomes from our real-world sample, then standardized the distribution to nullify drift, and ran it for an equivalent number of years a thousand times over. When the same MACD strategy was applied to the average of our 1,000 iterations the exposure was approximately equal but captured only 1.01% of returns relative to buy & hold from the same data (1.68% effectiveness).
In other words, the strategy was exponentially more effective when participants could see and react to the changing data than when they could not.
So then, why do we call technical analysis a “rudimentary form” of behavioral analysis? People typically rely on technical analysis as a way to bypass their own biases. When the decisions are made according to mathematical standards applied to chance, there is no place for your biases, but that doesn’t eliminate biases. Instead, it simply replaces your own with those of the herd following similar metrics… people don’t use technicals because they work, they work (when they work) because people use them.
