There is an intelligence test, where mice outperform humans.

It involves an experimental setup with two buttons – one red, and one green. One among these two buttons is the ‘correct’ button. There is a random 80% chance (4 out of 5) that this button is the green button, with a 20% chance of it being the red button. If a mouse pushes the correct button, it is rewarded with a crumb of food. If it pushes the wrong button, it receives a mild electric shock.

In such a setup, animals such as mice and pigeons quickly figure out that pushing the green button all the time was their best bet. They accepted the 20% penalty in return for an 80% reward.

In a similar setup, humans score a mere 68% – worse than these critters.

We humans also push the green button most of the time. However, we then observe patterns and try to predict when the red button offers the reward. Given that the setup is entirely random, our attempts at predicting a pattern are futile, and therefore, our performance is worse. Embarrassingly, members of our species persisted in this behaviour even after being told that the setup is completely random.

The side-effect of having an excellent instrument for pattern recognition is the tendency to see patterns everywhere, even when they don’t exist. In a random environment with low-validity, a ‘dumb’ strategy is sometimes the smarter one.

Now there’s a lesson we can learn from mice and pigeons.

Suggested reading: The trouble with humans