The Real Issue In Trying To Find PatternsDecember 8, 2012 12:07 pm
We all try to identify patterns, and apply them to predict the future. We do this everywhere, in each one of our actions. We basically ask, what has happened in the past? And then, we either say, the pattern will repeat, or we say, this time it would not repeat. Either ways, we try to look for patterns. It is probably just a built-in feature in human beings.
The other aspect to this, and we have written about this before, is that all of us have a tendency to over-react to recent events. It is obvious in many of our actions, starting with what we talk about all the way to where we spend or invest. But the reaction could be in two ways – we see the trend, and either think that the trend will break, or that the trend will continue. Why is this important? Turns out, this is the easiest way to classify people: take them to a Roulette table that has thrown three consecutive Reds. Do they bet on black or red?
But the biggest learning for us is this, and this is very important: the number of times Red turns up or Black turns up will become approximately equal only when the number of turns on the Roulette wheel is very high, running into Millions of turns..
For any of us to expect that over a small number of turns this probability will work is just being stupid.
In general, the published wisdom is for “millions of turns .. “, “in the long run ..” etc.
And that is the real issue with finding patterns. Think about it – we all believe that the sample of events we have at hand, is representative of the population always. And we try to look for patterns in the sample. And we take decisions based on that. In all fields. We are just being stupid.