Pattern Problems
Imagine participating in an experiment. You're shown several sequences of 20 flashing lights colored either green or red. Your job is to predict the color of the next flash of light, and you will receive a score on the accuracy of your predictions.
And you will be competing with pigeons.
Sample typical sequences: RGRGGGGGRGGGGRGGGGGG; RRGGGGGRGGRGGGGGGGGG.
The flashes are random but always with green flashing 80 percent of the time and red just 20 percent of the time.
Rats or pigeons, when rewarded for correctly guessing the right color, quickly learned the best strategy: green every time, guaranteeing an 80 percent success rate.
But humans didn't do nearly as well. Instead of picking green every time, they got caught up trying to guess the next red flash. On average, they scored only 68 percent instead of locking in the 80 percent by sticking with green. And they often performed worse the longer they tried, because they thought they’d discovered the "pattern" among purely random flashes.[i]
Humans are hardwired to search for patterns or causal relationships even when there is no such thing.
See the book for more patterns where none exist. On Amazon the whole chapter on patterns is available under “Look Inside.” Hint: Hot hand in pro basketball, reward and punishment, conspiracy theories.
[i] Jason Zweig, Your Money and Your Brain, 59.