DancesWithCamels

As I was doing some reading in the never-ending effort to increase my knowledge of war and politics, I happened to start analyzing the practical applications of Game Theory to current situations in the Middle-East. (As a warning to anyone interested in exploring Game Theory, it is heavy and tedious stuff. I apologize ahead of time if this writing reflects such.) Specifically, I focused in on the Game Theory problem commonly known as the “Prisoner’s Dilemma.”

For those who are unfamiliar with The Prisoner’s Dilemma, it is best explained like this: Two criminals are apprehended by the police (we’ll call them Criminal X and Criminal Y.) The authorities don’t have enough evidence to convict either man to a long prison sentence. Consequentially, the police isolate and interrogate each criminal separately. In an attempt to achieve a conviction, the police separately offer each criminal a deal.

The deal is this: if one of the convicts sells out his accomplice and the other refuses to cooperate with the police, the “rat” will walk free and his accomplice will serve a 20 year prison sentence. If both fail to cooperate with the police, then both will spend sometime in prison but will be released after 5 years. If both convicts cooperate with the police and rat the other out, then both will receive a 10 year prison sentence.

This leaves four possible results.

1. Criminal X rats out Convict Y. Criminal Y stays silent. Criminal X goes free. Criminal Y receives a 20 year sentence.
2. Criminal Y rats out Criminal X. Criminal X stays silent. Criminal Y goes free. Criminal X receives a 20 year sentence.
3. Both Criminal X and Y stay silent. Both Criminal X and Y receive a 5 year sentence.
4. Both Criminal X and Y rat out the other. Both Criminal X and Y receive a 10 year sentence.

Obviously, the dilemma for the prisoner is in what decision to make. Emotional variables such as loyalty and fear will influence their decision. Ultimately, someone will go to prison.

This exercise is interesting in and of itself, however, what really makes it fascinating (and more applicable) is when it is evolved into what is known as the “Iterated Prisoner’s Dilemma.” This form is the same as the original except that the “game” is repeated and repeated AND that the players have a memory of previous rounds. It is important to keep in mind that the Criminals are not necessarily competing against each other. They are simply trying to serve as little time in prison as possible. Though both Criminals must make decisions based on how they think the other will act, the scenario does not automatically condemn them both to long prison sentences. Consequentially, it is to be considered a “Non-Zero Sum” game.

This repetition of the process and the memory of the players offer the potential for retaliation (or cooperation) by either of the “Criminals.” (For instance, if Criminal Y stays silent but gets betrayed in Round 1, he may feel the desire to retaliate against Criminal X by ratting him out in Round 2.) Of course, as before the desire for each of the “Criminals” is to serve as little time in prison as possible. The individual strategies employed will vary based on any number of factors.

A study was conducted in the 1980’s with the purpose of developing the optimal strategy for the “iterated” version of the game. Numerous academics, theorists, and computer programmers participated and numerous strategies were tested. After testing, an analysis of the top scoring strategies revealed several interesting facts. To come out on top in the game, a strategy must include four key components.

1. It must not engage in a “First Strike.” That is, the player cannot “rat out” the other player before it is itself first the victim of such.
2. It must be ready to retaliate. That is, a player who is “sold-out” must be prepared to retaliate in kind on the next round. (Strategies which were unwilling to retaliate were taken advantage of by less pacifistic strategies.)
3. It must be willing to forgive. The strategy must be willing to occasionally forgive being “sold out’ by the other player. This means that if “sold out” the player must occasionally NOT retaliate even if it usually does. This is necessary to break a chain of retaliations which might occur. The inclusion of forgiveness is generally what separates a “Good” strategy from the “Best” strategy.
4. It must not be “envious.” Meaning, the strategy cannot have the objective of trying to score more than the other player. (A “No First Strike” strategy precludes this.)

Critical Point #1: A “Good” strategy requires a player to retaliate. It is better to always retaliate than to let the other player repeatedly take advantage of a pacifistic strategy, even if both parties selling the other out is less optimal than if both stayed silent.

Critical Point #2: The best strategy to be pursued by each player individually therefore contains these four elements. IF both players pursue this optimal strategy, then in the end both players will stay silent round after round. This means that even a selfish player, for completely selfish and utilitarian motivations, will be restrained and forgiving.

Critical Point #3: However, a vengeful player may act hatefully round after round despite the fact that it WILL result in less benefit than if it had showed emotional restraint.

Now let’s change the scenario to reflect a war and peace scenario. In this case we will use the Palestinians and the Israelis. We will replace the option of selling-out the other player with the generically titled “Acts of Aggression.” We will replace the option of staying silent with the generically titled “Following the Peace Plan.” Instead of prison sentences we will use a simple numeric scoring system. For the purpose of the exercise we will say that a point equals a measure of any combination of economic prosperity, political/religious legitimacy and/or physical security.

Thus our model will go like this:

If Israel “Follows the Peace Plan” and the Palestinians engage in “Acts of Aggression,” then the Israeli’s will receive 0 points and the Palestinians will receive 10 points.

If Israel engages in “Acts of Aggression” and the Palestinians “Follow the Peace Plan” then the Israelis receive 10 points and the Palestinians receive 0 points.

If both Israel and the Palestinians engage in “Acts of Aggression” then both receive 1 point.

If both Israel and the Palestinians “Follow the Peace Plan” then they both receive 5 points.

This means that IF both the Israelis and the Palestinians follow the best possible strategy, they will both continuously “Follow the Peace Plan” despite any hate or anger aimed towards the other. (See Critical Point #2)

The problem regarding the situation between the Israelis and the Palestinians (or the Sunnis and Shiites, etc.) is that there are far too many individuals who would rather act vengefully despite the decreased benefit of such action. An unwillingness to forgive means that both parties, though they may be following a “Good” strategy, (See Critical Point #1) they are not following the “Best” strategy.

This is about where the application of all this comes to an end. Despite the somewhat technical components of it all, this is still an overly simplistic explanation of the situation.

It seems clear that the course of action which would lead to the most benefit to both parties would be for both the Israelis and the Palestinians to cooperate together and develop (and stick to) a peace plan.

Tragically, I fear this will never happen. Despite the seeming rationality of both parties following the “Best” strategy, it remains elusive in the Palestinian-Israeli conflict. This is the region where the concept of an “Eye for an Eye” was invented. Tribal retribution is a way of life in much of the area. Culturally, forgiveness is in short supply. Based on my research and perspective, forgiveness is the key ingredient missing. With the addition of forgiveness, it might be possible to break the chain of conflict.

It is important to remember that in the hypothetical game, the cost of either side not following the “Best” strategy is only a loss of points.

In the real scenario, the cost is human lives…