In order to get around the world, we make predictions about what other people do. We guess that a friend will remember to make a reservation for a lunch date next week. We assume that a spouse will pick up a bottle of wine to bring to a friend's party. We expect that our children will choose the same item from the menu at a restaurant that they always do. These predictions work well enough. Often, our predictions are borne out.
Yet frustratingly, sometimes they are not. We make predictions based on our beliefs about other people's behavior, and sometimes people do not do what we expect. We come to rely on other people's predictability to the point where their failures to be predictable often lead us to wonder what is wrong with them.
As it turns out, though, unpredictability is a fundamental part of human nature.
Consider a simple case. Imagine we play a simple game. Every 10 second, I will flash a light, and then you have to press one of two buttons in front of you. If you pick the ‘right' button on that trial of the game, you get $5, and if you pick the wrong button, you get nothing. You will get to play the game for 20 minutes. There are 120 trials of the game in 20 minutes, so you could win as much as $600. Not bad for 20 minutes' work.
Now, the way the game is set up, the computer controlling the game picks a button randomly so that 70% of the time, the button on your left is the winner and 30% of the time, the button on the right is the winner.
What will you do in this game?
Chances are, by the last few minutes of the game, you will end up picking the button on the left about 70% of the time and the button on the right about 30% of the time. This behavior is called "probability matching" because the proportion of selections you make of each button is about equal to the probability that the button will be the winner.
Unfortunately for you, this strategy is not the best one you could choose. The best strategy you could pick is one called "maximizing." In this strategy, you should start by sampling the buttons to figure out which one is more likely to be the winner. After that, you should always pick that button. If you always pick the button that is rewarded more frequently, you'll win 70% of the time. If you pick the worse button 30% of the time, as you would with probability matching, you would probably lose most of the time that you picked that button, and so your overall payoff would be lower.
That means you'd be best off if your behavior was complete predictable based on which choice is currently the best one. Yet, your cognitive system is designed to make your behavior more unpredictable. On each trial of the game, there is only a probability that you'll pick the thing that looks like the best thing to do. That means that your cognitive system is willing to pay a price in the short term in order to be unpredictable. So, there must be some value in it.
Why are you unpredictable? There are lots of reasons, but let's just talk about two for now.
First, life is rarely like these simple games. In the game, I set up a payment schedule for the buttons, and it doesn't change. Most of the time, though, the state of the world does change over time. If you only shop at one store all the time, then you might miss out on a good bargain at a competing store. If you only go to one restaurant, you might not notice the slow steady decline in the quality of the food there, and lose out on the chance to eat at a new restaurant whose quality is improving.
That is, our world is a constant tradeoff between exploiting choices that have been good to us in the past, and exploring new options. If we only exploit, then we run the risk that we will not notice change in the world that make the option we are choosing worse than it was or make other options in the world better than they had been when we first explored the world.
Second, if our behavior is completely predictable, then it makes it easier for other people to take advantage of us. Now, in the modern world, there may not be lots of people lurking to take advantage of us. But in the evolutionary environment in which we (and basically all other species on Earth) evolved, there was always a danger that someone (or some thing) would be out there taking advantage of us. So, some unpredictability has served us well., and those mechanisms that make us unpredictable are still a fundamental part of our cognitive system.
So, the next time that someone does something unpredictable, remember that they are wired to do it. And remember that if they weren't unpredictable, they would be losing out in the search for the best things that life has to offer.