According to Game Theory, Chimpanzees are better than humans.
In a January 2013 TED talk, Colin Camerer talked about how the brain strategizes. Specifically, he talked about research that analyzed what parts of the brain are active in tests that use game theory. The research is based around Cognitive Hierarchy Theory, (CHT), which is essentially a theory that attempts to describe thought processes in strategic games. CHT assumes that the players base their decisions on what they think the other players are likely to do. For instance, in one study that Camerer discussed, two players were involved in a bargaining game where one player was “informed” and knows the amount of money being split, while the other player was “uninformed” about the total amount of money, which ranged from $1 to $6. Just like the two-way bargains that were discussed in class, if the “uninformed” person agrees to the bargain, he gets the proposed amount of money and the “informed” player gets the rest, but if he doesn’t agree, then neither person gets any money. The two players only have ten seconds to make a bargain. It turns out that the “uninformed” player is most likely to disagree if the amount being presented to them in the bargain is small, while they will agree much more regularly if the amount is larger. The decreasing rate of disagreement with increasing monetary offers appears to be approximately linear based on the graph presented. (It’s also interesting to note that people who played face-to-face vs over a computer were much more likely to split the money evenly every time).
Camerer also discussed a study where people were asked to choose a number between 0 and 100. All of the submitted numbers would be averaged, and whomever was closest to 2/3 of the average would win. People who only thought through one step of this problem chose values around 33, assuming that the average of randomly submitted numbers between 1 and 100 would be 50, and 2/3 of 50 is 33. Some people took this one step further, and proposed around 22, which is 2/3 of 33. Another group of people figured out the Nash equilibrium for the problem by realizing that if you keep calculating steps of the averages, everyone wants to be below everyone else and so you arrive at 0. Thus, there were three nodes in the resulting graph of submitted numbers: one at 33, one at 22, and one at 0. (It turns out that the actual average was 23.07). Thus, it is interesting that the “smartest” people who calculated the Nash equilibrium were not the most accurate, but instead that people who calculated two steps of the problem seemed to be. Camerer was interested in why this is and what makes people think in such a way.
After talking briefly about how fMRI’s show how people think depending on whether they believe that they are playing another person versus a human, Camerer moved on to the two-player bargaining with an “uninformed” and an “informed” player, (discussed earlier). It turns out that the player’s brains are very closely synched up when the players will agree, with the “uninformed” player’s brain indicating activity slightly before the “informed” player while the brains had a much higher level of activity, with much of it starting with in the“informed” player’s brain, if the two disagreed.
Finally, Camerer took his discussion of how people think a step further, and talked about a study that involved a game between either two humans or two chimpanzees. In this game, each player would be randomly selecting a box, right or left. One of the players won if both players picked the same box, (both chose right or both chose left), while the other won if the boxes were mixed, (one chose left and one chose right). When the payoff for choosing either box was the same, the chimpanzees and humans each tended to pick right and left with about the same probability, which is a Nash equilibrium in this case. However, if the payoff for choosing one box over the other and winning was increased for the player who wins by matching, (for example, if choosing left and winning has a higher payoff than choosing right and winning or vice versa), the chimpanzees tended to alter the frequencies of their box selections to align with the new Nash equilibrium while the humans stayed around the 50/50 region of right/left box selection. Chimpanzees actually got within 0.01 of the new Nash equilibrium, “closer than any other species we’ve observed,” according to Camerer.
Thus, in this TED Talk, Camerer showed that there is evidence that the brain has early warning signs of disagreements in bargains and that chimpanzees are better competitors than humans.