Can Past Patterns Be Used to Predict the Future?

Some researchers who have accepted that stock prices do not follow a random walk still do not accept the validity of technical analysts. These opponents agree that there may be patterns that can be fitted to stock price movement after the fact, but they argue that these past patterns cannot be used to predict the future. In other words, past patterns cannot be exploited to gain above-average returns. There are two major reasons why this group of opponents, especially academics, has drawn these conclusions.

First, though there may be some underlying patterns, markets are constantly being affected by new information. This new information causes enough variability in the underlying pattern that any knowledge of the underlying pattern will not be enough to exploit the knowledge for profit. For example, a recurring business cycle is a well-known and accepted economic phenomenon but not a predictable harmonic. The economy experiences expansionary periods followed by recessionary periods repeatedly. Therefore, we can expect cycles of expansions and contractions in the future. However, each of these business cycles is unique; the cycles vary in length and intensity. Thus, acknowledgment of a recurring cycle cannot be equated with the ability to predict the timing of an expansion or the intensity of a recession.