Want to keep learning?

This content is taken from the University of Groningen's online course, Decision Making in a Complex and Uncertain World. Join the course to learn more.

Skip to 0 minutes and 10 seconds In the previous lectures, we have seen that all organisms, including humans, are confronted with a complex and uncertain world. All organisms, including humans, have to take decisions under such circumstances, and the brain processes underlying these decisions have, at least partly, been shaped by natural selection.

Skip to 0 minutes and 31 seconds In group living animals, including humans, complexity and uncertainty is, to a large extent, caused by the social environment. Organisms have to take decisions, but where the decision turns out to be good or bad strongly depends on the decisions of others. The fact that, in the social context, decision making should reflect the decision of others, it’s a whole level of complexity.

Skip to 1 minute and 1 second The implications of this feedback between own decisions and the decisions of others are the topic of this lecture.

Skip to 1 minute and 10 seconds Decision making in a social setting can be analysed with the methods of game theory– a theory trying to find out what kind of behaviour is optimal in situations of conflict and cooperation. This theory was first developed for decision making in economics and the political sciences, but has been proven to be more successful in biology and the social sciences.

Skip to 1 minute and 37 seconds The name “game theory” derives from the fact that the real world conflict situation is modelled in a similar manner, as one would model a parlour game, like chess or poker. Accordingly, the participants in a conflict are called players. Each player has a set of strategies available that determine the moves to be made in any conceivable decision situation. In the end, each player receives a payoff, which can be positively or negatively correlated with the payoffs of other players. Typically, the interests of the players are partly overlapping and partly conflicting with each other.

Skip to 2 minutes and 18 seconds But economic game theory tends to assume that the interacting agents are, to a large extent, rational, whatever this means. Evolutionary game theory assumes that natural selection will single out strategies that are “evolutionarily stable”, in the sense that no mutant can outsmart others in a population where all individuals are playing an “evolutionary stable” strategy.

Skip to 2 minutes and 43 seconds In earlier lectures, we have already seen that the children’s game like Rock-Scissors-Paper may represent a real world situation, like competition among algae. In this lecture, we will focus on a game that tells us something about the evolution of cooperation.

Skip to 3 minutes and 0 seconds By definition, cooperation is the joint behaviour of interaction partners that is beneficial for all individuals involved. Yet, it is not straightforward that cooperation will, indeed, evolve. The problem is that the cooperation can often be exploited by free riders who reap the benefits of cooperation without contributing to the costs.

Skip to 3 minutes and 26 seconds The prototype example for this problem is a so-called Prisoner’s Dilemma game, which was first started in the political sciences. No other game has attracted that much attention, and it still inspires hundreds of scientific publications every year. In the reading material, we will explain why a game exemplifying the evolution of cooperation has such a strange name. We will also show that the Prisoner’s Dilemma game has various biological applications.

Skip to 3 minutes and 59 seconds The Prisoner’s Dilemma game has two strategies, called cooperate and defect. Cooperation yields a benefit B, but it comes at a cost C. The benefit B is assumed to be larger than the cost C. The conundrum is that the benefit B is obtained when everyone’s interaction partner cooperates, while the cost C only has to be paid by the cooperator. l What now is the dilemma? Well, mutual cooperation yields B minus C for both players, which is positive. This is better than the payoff of mutual defection, which is zero. Still, it is better to defect under all circumstances. If the other player cooperates, a defector receives B. What is larger than B minus C?

Skip to 4 minutes and 52 seconds And if the other player defects, the defector receives zero, which is larger than minus C. For this reason, the strategy defect is the only rational strategy when humans are caught in the Prisoner’s Dilemma type of situation, and defection would also be the “evolutionarily stable” strategy in a biological context, even if mutual cooperation is much preferred to mutual defection. This is our first take-home message. In a social context, decisions that are mutually favourable for all parties will often not evolve since that they can be exploited by alternative strategies. Evolution will not lead to strategies maximising mutual payoffs, but to evolutionary stable strategies– that is, to strategies that cannot be outsmarted by alternatives.

Skip to 5 minutes and 48 seconds Even in the Prisoner’s Dilemma context, corporation can evolve if the same individuals are playing a Prisoner’s Dilemma game repeatedly. In such a situation, unconditional cooperators would still be out-competed by defectors. However, now smarter cooperation strategies are available. By making their own cooperation conditional on the cooperativeness of the other player, these strategies are better protected against exploitation.

Skip to 6 minutes and 19 seconds The prototype example for such a responsive strategy is Tit-for-Tat. Tit-for-Tat players start with cooperation and continue to cooperate as long as the other player cooperates, however, they switch to defection when having been the victim of defection in the previous round of the game. In the presence of Tit-for-Tat, the strategy “always defect” is no longer a dominant strategy. A defector can exploit a Tit-for-Tat player only a single time, and from that point onward he will receive zero. The accumulated payoff of a defector is therefore quite small, and typically smaller than the payoff received by Tit-for-Tat player against another Tit-for-Tat player.

Skip to 7 minutes and 2 seconds A strategy like Tit-for-Tat can, therefore, out-compete a strategy like “always defect” and still end up in the favourable situation of mutual cooperation. In the population of Tit-for-Tat players, nobody will be the first to defect and, hence, everybody will cooperate all the time. This is my second take-home message. In the social context one should expect evolution of strategies that make their behaviour dependent on the local circumstances, and in particular, on the behaviour of others. The availability of such conditional strategies can strongly affect the outcome of social evolution.

Skip to 7 minutes and 42 seconds But there’s a problem. It may be difficult to find out what kind of conditional strategy is being employed by the participants in the social interaction. This can, again, be illustrated by Tit-for-Tat. Take a population of Tit-for-Tat players– everybody will cooperate and, hence, the population shows exactly the same behaviour as a population of individuals prepared to cooperate unconditionally, yet the two populations have very different stability properties. While unconditional cooperation can easily be exploited, this is not the case for Tit-for-Tat since Tit-for-Tat individuals are willing to strike back.

Skip to 8 minutes and 23 seconds The third take-home message, therefore, is, in social interactions, not all aspects of a strategy are directly observable. In fact, crucial aspects determining the success of a strategy are often hidden below the surface.

Skip to 8 minutes and 41 seconds Tit-for-Tat is not the only good strategy in the repeated Prisoner’s Dilemma game. In fact, Tit-for-Tat is quite vulnerable to small errors. Suppose, for example, that a Tit-for-Tat player erroneously defects in one round. If the other player plays Tit-for-Tat as well, this accidental error will induce retaliation, which in the next round will be punished by the first player, and so on. A more forgiving strategy, like Tit-for-2-Tats, may therefore, be a superior strategy. In fact, there are many potential success strategies in a repeated Prisoner’s Dilemma game, and not all of these strategies eventually end up in neutral cooperation.

Skip to 9 minutes and 23 seconds This holds for virtually all repeated games. There are infinitely many unexploitable strategies, and interaction among these strategies can lead to a broad spectrum of possible outcomes.

Skip to 9 minutes and 35 seconds My fourth take-home message, therefore, is– in social interactions, there is typically a huge variety of possible success strategies. The problem is to find out which of these strategies is being employed by one’s interaction partners.

Skip to 9 minutes and 53 seconds The fact that the simple game like the repeated Prisoner’s Dilemma has a multitude of potential success strategies has important implications for the evolutionary dynamics of such strategies. Simulations clearly show the signature of non-equilibrium dynamics. In an earlier lecture, we have seen that non-equilibrium systems have very different properties than equilibrium systems. Hence, the fifth take-home message is– even simple interactions with few participants can have very complex dynamics. Game theoretical equilibrium arguments should be treated with extreme care.

Skip to 10 minutes and 31 seconds In this lecture, we have seen that an extremely simple model for social interactions has already a very rich structure, and that decision making in the social context is more complicated than one might think.

Skip to 10 minutes and 44 seconds Time does not permit to discuss other seemingly strange aspects of social decision making, but I would like to close with one of these aspects, which is explained in detail in the reading material. There we discuss the effects of giving one of the players additional information, or additional strategic options. One might think that the better informed or better equipped player is in a superior strategic position. As you will see, this is not necessarily the case. It is easy to construct situations where player having additional information or additional options is actually in a strategic disadvantage, leading to adverse outcome. The general conclusion of this lecture, therefore, is decision making in a social context has many surprising, and often counter-intuitive features.

Skip to 11 minutes and 34 seconds The physical world may already be complex and uncertain, but it may well be that the social environment induces even more complexity and uncertainty to lives of all organisms, including humans.

Introduction to social evolution

This video introduces the concept of social evolution. Our environment is to a large extent complex and uncertain due to its social aspects. Whether a decision is good or bad strongly depends on the decisions of others. To analyse this, we can use game theory.

If you find the ‘tit-for-tat’ game difficult to understand after having watched the video you can have a look at this Wikipedia article.

Share this video:

This video is from the free online course:

Decision Making in a Complex and Uncertain World

University of Groningen

Get a taste of this course

Find out what this course is like by previewing some of the course steps before you join: