Trust as Explained by Game Theory

This was an interesting page/exercise sent to me via Twitter applying the concepts of game theory to the generation and maintenance of trust.

People no longer trust each other. Why? And how can we fix it? An interactive guide to the game theory of trust: http://ncase.me/trust/

It takes like 20-30 minutes to complete. At first I was turned off a little by the arbitrary constraints of the game, but they end up dealing with that issue later on -- so patience pays off! I've seen these models before, but it was interesting to apply it more directly to trust. Also, I hadn't seen the addition of mistakes/misunderstandings into the model before too. That has already changed the way I view others and the world. For instance (this might not make sense until you do the exercise), a friend of mine recently had an Amazon package fail to be delivered. She assumed that it was some shady neighbors stealing the package and was going to stop having any packages delivered, even though she has had like 20 successful package deliveries so far. I encouraged her to keep trying until she has another package go missing, just in case there was a mistake or other one off occurrence that shouldn't necessarily change her game playing strategy. It's a risky strategy maybe, but in her case she has no other convenient alternative for package delivery.

Without really remembering, I had applied essentially the "Diamond Rule" to this game. I think this worked ok (and probably works better with actual people than bots?), but it is true that in a situation in which there is a mistake, it can also compound a mistake into a global loss.

There's that phrase "fool me once, shame on you, fool me twice shame on me". But this game suggests a more optimal rule, when mistakes are factored in: "Fool me once, ok, I take it on the chin. Fool me twice, shame on you with punishment."

Ethical sociopaths?

Can sociopaths be ethical? A certain type of ethical, certainly. But before we talk about potential sociopathic limitations, this NY Times article ("In Life and Business Learning to be Ethical") about the issues with ethics that almost all of humanity shares:

The problem, research shows, is that how we think we’re going to act when faced with a moral decision and how we really do act are often vastly different.

Here’s just one of many examples from an experiment at Northeastern University: Subjects were told they should flip a coin to see who should do certain tasks. One task is long and laborious; the other is short and fun.

The participant flips the coin in private (though secretly watched by video cameras), said David DeSteno, a professor of psychology at Northeastern who conducted the experiment. Only 10 percent of them did it honestly. The others didn’t flip at all, or kept flipping until the coin came up the way they wanted.
***
[W]e need to be more aware of the ways we fool ourselves. We have to learn how to avoid subconsciously turning our backs when faced with a moral dilemma. And then we must be taught how to challenge people appropriately in those situations.

“When people predict how they’re going to act in a given situation, the ‘should’ self dominates — we should be fair, we should be generous, we should assert our values,” said Ann E. Tenbrunsel, a professor of business ethics at the University of Notre Dame who is involved in the EthicalSystems website. “But when the time for action comes, the ‘want’ self dominates” — I don’t want to look like a fool, I don’t want to be punished.

“Our survival instinct is to want to be liked and to be included,” said Brooke Deterline, chief executive of Courageous Leadership, a consulting firm that offers workshops and programs on dealing with ethical situations. “We don’t willfully do bad things, but when we’re under threat our initial instinct is to downplay or ignore problematic situations.”

Sociopaths may not have the same set of ethics (or issues implementing ethics -- our survival instinct is not so much to be liked and included), but it's also possible for sociopaths to have a personal preference about how they wish to act, even if it is just a personal aesthetic as opposed to be a moral code. Webster defines ethics as many things, including "a guiding philosophy". Maybe for sociopaths that would look something more like utilitarianism or "the diamond rule", as opposed to a saint's altruism and golden rule, but even criminals have codes.

At the heart of any choice to ascribe to a set of ethics, whether empath or sociopath, is a belief that your choices matter -- that you and others around you are affected by everything you choose to do. You don't have to believe in right and wrong to understand that you are what you eat. And I wouldn't want life any other way. What would be the point of making choices if they didn't matter? And if you believe your choices matter, it's only natural to ascribe to some sort of "guiding philosophy" about how to make those choices. So yes, sociopaths can and are ethical. Could sociopaths ever be considered more ethical than empaths? 

Traders > psychopaths

As a sometime dabbler in the stock market, I found amusing this Der Spiegel article about traders being not necessarily more ruthless than psychopaths, but definitely more pointlessly vindictive:

According to a new study at the University of St. Gallen seen by SPIEGEL, . . . stockbrokers' behavior is more reckless and manipulative than that of psychopaths. Researchers at the Swiss research university measured the readiness to cooperate and the egotism of 28 professional traders who took part in computer simulations and intelligence tests. The results, compared with the behavior of psychopaths, exceeded the expectations of the study's co-authors, forensic expert Pascal Scherrer, and Thomas Noll, a lead administrator at the Pöschwies prison north of Zürich.

Appetite for Destruction

"Naturally one can't characterize the traders as deranged," Noll told SPIEGEL. "But for example, they behaved more egotistically and were more willing to take risks than a group of psychopaths who took the same test."

Particularly shocking for Noll was the fact that the bankers weren't aiming for higher winnings than their comparison group. Instead they were more interested in achieving a competitive advantage. Instead of taking a sober and businesslike approach to reaching the highest profit, "it was most important to the traders to get more than their opponents," Noll explained. "And they spent a lot of energy trying to damage their opponents."

Using a metaphor to describe the behavior, Noll said the stockbrokers behaved as though their neighbor had the same car, "and they took after it with a baseball bat so they could look better themselves."

The researchers were unable to explain this penchant for destruction, they said.

As continued here:

Faced with a hypothetical choice between co-operating for everyone’s benefit and getting a predictable reward or cheating and possibly getting more for themselves, traders were more likely than psychopaths to cheat, said Noll.

As a result, psychopaths, who broke the rules occasionally, won the most, ordinary people, who almost always played by the rules and who co-operated, came in second, while traders, who didn’t care how their actions affected anyone else, cheated the most and won the least.

Sounds like a violation of the Diamond Rule to me.

The Diamond Rule

This is a really interesting explanation of the application of game theory principles to ethics. I found this to be particularly useful in categorizing my own behavior and those of others:

The first strategy is called “The Golden Rule.” This strategy basically represents the idea to “Do to others, what you would have them do to you.” So if you were to apply this to the prisoners dilemma, you would remain silent every time the game was played.

Another strategy available is called “The Silver Rule” which is to “Do not do to them, what you would not have them do to you” which is a variation of the golden rule, but in the case of the prisoner’s dilemma, would also mean that you would remain silent every time the game was played.

Another strategy available is called “The Bronze Rule” which is to “Do to them, as they do to you” or as some would say “an eye for an eye” which, in the case of the prisoner’s dilemma, would mean that you would remain silent, if the other prisoner remained silent the last time the game was played, or if they were to testify against you, then you would do the same to them the next time around.

Another strategy available is called “The Iron Rule” which is to “Do it to them, before they do it to you” which, in the case of the prisoner’s dilemma, would mean that you would testify against the other prisoner every time the game was played.

These different strategies were all tested in multiple computer simulated and in real world experiments, and it was found that actually none of these strategies were superior for the long term benefit of both players.

It was found that a completely different strategy was found to be superior, which has been called the “Tit for Tat” strategy, and which I have labeled as “The Diamond Rule.”

This strategy basically states that both players should start out with the Golden Rule/Silver Rule strategy of cooperation with the other player, or in other words, to remain silent each time the game is played. If at any time the other player is to defect, or to testify against you, you are then to “punish” them, by also testifying against them the next round of the game, but then after that round, to forgive them, and to return to the Golden/Silver Rule strategy of cooperation.

In effect, the Diamond Rule states that you should cooperate with others, but if at any time another person is to hurt you, then you must withdraw your cooperation and get justice until justice has been fully dealt, and to then “forgive and forget” and be open to cooperating with them once again.

So that is how the Nash Equilibrium and Game Theory is most readily and easily applied to ethics. In effect, the Nash Equilibrium of Game Theory is a mathematical proof for why things like ethics, altruism, justice and forgiveness are actually most beneficial for every member of the group in which they exist. A mathematical justification for moral and ethical behavior.

I'm not surprised that the Diamond Rule was the most efficient. And longtime readers of the blog shouldn't be surprised to see me assert that I think sociopaths are probably the most able to pull the Diamond Rule. First, you have to be able to punish someone, harsh enough to make an impact and fast enough so you can both quickly move on to the forgiveness stage. Sociopaths are impulsive and ruthless, they would easily be able to pull this part off. Second, you would have to completely forgive them and go back to the way things were before. Normal people hold grudges, but for sociopaths something like this isn't personal, just business. Once you have knocked them around for their misbehavior, if they have any use to the sociopath it would be right back to being chummy. In fact, the Diamond Rule is pretty much my default. I hate the Golden Rule, I think it's idiotic and narcissistically inefficient. I do the Silver Rule, but will take them down if necessary (Diamond Rule). The Bronze Rule is petty and others centered. The Iron Rule is just too inflexible for modern society.

In other news, I'm sort of fascinated and flummoxed by the author of the blog (linked above), "Broderick Boyd" of "The Broderick Boyd University." I can't tell whether he is trying to become a motivational speaker, is a raging narcissist, or what, but I'm curious enough to follow him on Twitter.