The game of trust

How do people establish trust under the most adverse circumstances? How can sworn enemies get into a position that it makes sense for them to help each other.

In game theory, this is the surprising phenomenon that the prisoner’s dilemma aims to explain.

Imagine two suspects held by the police who don’t have enough evidence against them. Their choices (game moves) are to confess or stay silent. If they help each other by staying silent, they will both get a positive pay-off: they will be free. If one confesses, and therefore betrays the other, the second one will go to prison for a longer time than the first one. If they both confess, they will both go to prison.

The best strategy to reduce potential individual losses is to confess, but unfortunately it leads to a bad outcome for all.

The catch

The outcomes are very different, though, if the game is played repeatedly with the same players, and they let their decisions (their game moves) depend on the history of moves. In that case, cooperation is a very likely outcome, as we will see.

In his book “The Evolution of Cooperation” (1984), Robert Axelrod gives three very different stories on the outcomes of this scenario.

The first account takes us back to World War I trench fighting, where soldiers would often be within shouting distance of the enemy. That frontline hardly moved in years, as it turned out that soldiers quickly learned that every assault would be answered directly by enemy fire, and that every break in fighting would lead to less shooting from the other side. There is even a famous story of the two sides coming out of the trenches at Christmas time to toss around a football. This behavior was so pervasive that officers took drastic measures to break it up, for example by rotating soldiers over different locations.

The second account comes from a tournament that Axelrod organized. He invited other game theorists to submit a computer program that embodies a specific game strategy. In the tournament these were pitted against each other. The winner of the tournament was a very simple strategy: ’tit for tat’. Start by being nice, and then copy the opponents previous move. This outperformed much more complicated strategies, and typically leads to the highest total pay-off.

In the third account, Axelrod argues what essential attributes a winning strategy should have. They are that a strategy should be:

  • Nice: start out by showing good intentions, not by defecting.
  • Provocable: bad behavior should be retaliated, if not it can be exploited.
  • Forgiving: the strategy should not get into a ‘vendetta’ of continuous revenge and counter revenge.
  • Non-envious: the strategy should not strive to get more out of the game than an opponent.
  • Clear: make it easy for others to cooperate with you.

Of course, there are some assumptions built into the argument, for example on what types of other players there are in the tournament.

A fascinating side of the book is that it illustrates how cooperation can evolve in three very different ways that still lead to the same conclusion.

Trust evolves

When we talk about value, power, and risk, we also talk about collaboration and trust. Axelrod’s analysis shows us how we can analyze trust and trust relations, and what it takes for trust to be beneficial to both parties. In contrast to what some of us may have been told at kindergarten, being nice, forgiving and non-envious is not enough. You also need credible retaliation.

Who is playing?

The digital infrastructure world has many actors, as we discuss elsewhere. For example, A person using a cloud email system is an actor, and so is the email provider.

The degree to which they trust each other is often derived from the history of interactions. It is like a restaurant. If the food was bad, or the staff messed up, you are not coming back soon. Similar with a SaaS provider: bad service means you are thinking of going away.

In an enterprise, this may work a little different, but the same principles apply. If a service provider does not deliver according to the Service Level Agreement (SLA) there will be remedies to pay, or at least a stern conversation between the enterprise’s SLA manager and the provider’s account manager.

Tit for tat does not really work between a single individual and a large provider. If the provider does not serve the individual very well, it is a huge disservice to the consumer. But an individual cancelling their subscription is not making much of an impression in relation to the large population that the provider serves.

This asymmetry in power can be very frustrating to the ‘weaker’ party. And if the individual is part of an enterprise that holds the contract, they do not even have the option of stepping out.

Conversely, if the experience and power of enough individuals is combined, they can make a significant impact.

I worked for an internet service provider (ISP) in the early days. Back then, they were not all that good. Hence, consumer organizations (advocacy groups, associations) were compiling rankings of ISPs, and consumers actively searched for the better providers.

I reverse engineered how that ranking came to be, and helped the ISP understand where they needed to invest in quality. They then quickly rose to the #2 spot on the list, and accelerated their growth. By restoring the power symmetry in this way, the end result is better for both sides: good service and good market share.