Fisher Investments Press: 20/20 Money: Forecasting, Part 2—Recognizing Patterns

The four most expensive words in the English language are, "This time it's different."
    — Sir John Templeton

The fact that people will be full of greed, fear or folly is predictable. The sequence is not predictable.
    — Warren Buffett

Last chapter, Fisher Investments Press author Michael Hanson looked at methods of market forecasting using probability. In this chapter, we’re going to study the other side of forecasting—pattern recognition.

In Chapter 5 we observed CEASs sometimes show repeating and recognizable patterns. For instance, it may not be possible to know exactly which traits a strand of DNA will produce in a person, but we do know it'’s overwhelmingly likely to produce a human with two eyes, two ears, a nose, a mouth, and so on. We can even produce fairly reliable probabilities on eye, skin, and hair color, height, sex, and so on. Those are recognizable and probable patterns.

Or it might be impossible to predict the exact weather on a given day, but the earth’s meteorological systems are consistent enough that we can know in a general sense what the weather will be like. Those are demonstrable patterns of the larger system and reasonably predictable. In both examples, there is a great deal of randomness, or noise, but in that noise there are patterns.

Many believe this is how stock markets work. For instance, over the long-term (where there is less noise in the data than daily or monthly observations), stock markets tend to go up over time, tied to the profitability of the underlying companies they track (like we saw in Chapter 5). Fisher Investments Press author Michael Hanson believes that's been true for as long as markets have existed. This pattern appears to defy randomness. This chapter will outline some of the most prevalent patterns and drivers of stock prices.