How being a bit unreliable might be an advantage, new study
While being unreliable in relationships or at work is usually frowned upon, new research suggests these individuals might be doing what helps them thrive and survive.
A new game-theory study published in the journal ‘Proceedings of the National Academy of Sciences’ (PNAS) indicates evolutionary selection does not lead to flawless rationality.
Instead, a certain level of strategic unpredictability protects people from being exploited or overworked by their peers.
“If you act with perfect rationality, you become completely predictable, you give others a perfect roadmap to exploit you.”
Co-author of the paper, Christian Hilbe, Professor of game theory and evolutionary dynamics at IT:U.
People make strategic decisions all the time, but the outcome depends not only on what one person does, but also how others react to it, he says.
By being unreliable, or introducing “strategic noise”, those individuals force others to “feel compelled to accommodate them. They feel they cannot rely on them to do the best thing, so they naturally adapt to those people. In doing so, they do those people a favor”, says IT:U Professor Hilbe.
This dynamic happens everywhere, from classrooms to workplaces and relationships, such as people doing the work of their co-worker or co-student to speed things up, to the same people always doing the dishes.
Typical office scene: who is doing the dishes?
‘Making you do all the work’.
At the heart of it is a principle called ‘outcome sensitivity’, or how people respond to the perceived success of a behavioral strategy.
More sensitive people are more likely to use strategies they perceive as successful, while less sensitive ones learn in a more unpredictable way, making their behaviour more erratic.
The study’s scientists from IT:U, the University of Amsterdam and the Max Planck Institute have simulated people who played different economic games and recorded their behavior.
They have found that lower sensitivity can often pay off, Professor Hilbe says.
He gives the example of two university lecturers running a course together, with one offering exceptional student service and replying to requests at once.
The other one is a bit more unreliable or slower to respond, “but you really want to offer a good service to the students. So, you learn to just do it yourself. By being unreliable myself, I implicitly make you do all the work, even if you do not really want to do it”, says Professor Hilbe.
This model gives a precise explanation of why this type of behavior happens in the first place, the researchers say.
Surprising result
Professor Hilbe says the result is surprising as “naturally, the expectation would be that we always evolve towards higher rationality because it helps”, he says, “so, one would have expected that in the end we make super-rational decisions”.
This insight could also prove useful for the invention of future artificial intelligence, where over-rationality could be disadvantageous, while adding some more human-like or erratic manners might just be the key to success.