Random Walk Theory maintains that market movements are random, that stock price changes cannot be predicted by studying past prices. Random walk theorists compare historical price data to a series of coin tosses. They argue that since the outcome of a coin toss is not dependent on the results of earlier tosses, historical coin-toss data are of no predictive value whatever. If the analogy holds, then past price data are also independent, unlinked, and of no predictive value.
Random walk arguments fail to recognize that traders act with purpose. Since price activity is the result of traders acting out intentions, aggregate intention is often discernible in the behavior of price and volume. Since knowledge of intent has predictive value, markets are predictable.
Random walk supporters maintain that markets are moved by news. By definition, news is random, cannot be anticipated (or it would not be news), and therefore price movement is random and cannot be anticipated. They fail to take into account that the flow of price and volume activity is also news. Traders respond to that news just as they respond to other sorts of news, and so forge causal links between past price-volume data and future price movement. Price-volume data are causally linked because traders link them.
Random events can and do perturb prices unexpectedly, and the incessant clash of buyers and sellers often produces more noise than music. Nevertheless, from time to time the purposes of buyers and sellers collect sufficiently to produce a signal strong enough to be heard through the din.
The artful student is able to detect and exploit these opportunities, which present themselves daily.
In recent years, computerized trading systems have become popular. They range from very simple to very complex, from inexpensive to costly. Such systems, it is claimed, are emotion-free and therefore not subject to the destructive influences of fear and greed. Most have been back-tested over many years' data.
Computers are an important, perhaps vital, aid to trading. But there is no black-box trading system which will produce long-term, consistent profit.
Because markets are constantly changing, systems fail sooner or later. Markets are games in which the rules of play are freely and spontaneously created by the players. As the rules change, the nature of play changes. And change is constant.
Imagine a football game undergoing gradual transformation to a game of rugby, and then to a game of soccer. Someone trying to model the game would find it necessary constantly to recalibrate his model.
To work, systems require consistent, rule-governed behavior. But from bull to bear markets, from active to dull markets, what it takes to win is constantly changing. What works today may not work as well or at all tomorrow. Any system based on fixed rules will fail because it is static, but the market is not.
The premise of system builders and users is that markets are amenable to mathematical description, as are physical phenomena such as the diffusion of a gas or the cycles of planets. But the analogy is wrong. Markets are a form of human behavior, driven by human purposes, emotions, and needs. Markets are not fixed systems, but as elusive and adaptive as the humans who create them.
Buyers of systems are searching for one or both of two benefits: 1) quick profits without the drudge and expense of study and practice; and/or 2) protection against emotional trading decisions. Any mechanical trading system is a crutch, propping the trader against his own weaknesses. Such devices foster dependence. Those who rely on systems not only hinder their educations but also postpone the time when they will stand on their own feet.
Systems may model one aspect or another of market behavior, but the only computer capable of registering and processing the nuanced, many-faceted nature of any human activity is the one that sits between your shoulders. The single best system remains the well-tutored mind.