Clarissa Carrillo
Patricia Olachia
Main Points
1. Prisoner's Dilemma: a gambling game that has two card choices for
each player. They are called COOPERATE and DEFECT.
2. There are other versions of the game and each one has a strategy.
3. The overall most successful strategy was Tit for Tat.
4. Dawkins defines the terms "nice," "forgiving," and robust.
5. In order for a strategy to succeed, it must be able to do well in a
climate dominated by copies of itself.
6. Inhabitants find themselves clustered together with individuals who
resemble them.
7. In order for Tit for Tat to be most successful, it must always be
forgiving and never be envious
Summary: Nice Guys Finish First
In chapter 12 entitled Nice Guys Finish First, Dawkins begins by defining a nice guy according to Darwinian terms. He states that it is an individual that assists other members of its species, at its own expense, to pass their genes on to the next generation. According to Dawkins, it would seem that nice guys, therefore, become extinct. He proves in this chapter, however, that they actually succeed and finish first.
Since Robert Trivers' 'reciprocal altruism' principle, an American political scientist named Robert Axelrod took this principle in a different direction. He became fascinated by a simple gambling game called Prisoner's Dilemma. There are two possible cards that can be played by each player. The dilemma is that one player does not know what card their opponent will choose. The two choices are Cooperate and Defect. There are four possible combinations with these two choices. Each combination has a repercussion. I will briefly try to explain. Dawkins uses the example of two prisoners. (Hence 'prisoner's dilemma). The currency is not money, but prison sentences. Two men, Peterson and Moriarty, are in jail. Each is asked to betray his colleague (defect) by turning evidence against him. The outcome depends on what each prisoner does, and neither knows what the other has done. If Peterson blames Moriarty, and he cooperates, Peterson is free. If both betray each other, they both are convicted, but with a light sentence for cooperation. If both cooperate (with each other), they receive a small sentence because there is no evidence against either. In this case it would be perfect for both to cooperate, but neither has a choice but to betray each other and therefor suffer the consequences. There is a way out of this dilemma. There isn't, however, ways of ensuring trust. The best thing to do in this case, is to defect.
Dawkins discusses another version of the game called Iterated or Repeated Prisoner's Dilemma. This game is similar to the Prisoner's Dilemma, except it is repeated an infinite number of times with the same players. In this game, the successive rounds give us the opportunity to build up trust or mistrust, to reciprocate or placate, forgive or avenge. The important point is that both players can win at the expense of the banker instead of each other. If both players cooperate, then they win everything. If both players defect, then they each loose money to the bank. If one defects and the other cooperates, then they end up with intermediate sum of money.
Dawkins begins a section on different strategies and comes up with the best one. His first strategy is called Tit for Tat. This begins by cooperating on the first move and then copying the previous move of the opponent. If one player cooperates, then the other will copy that move and so on until they have both reached the benchmark. A second strategy is called Naive Prober. This strategy is similar to Tit for Tat, except that it randomly throws a defect. If these two strategies should play against each other, Tit for Tat will cooperate and Naive Prober will defect, therefore making Tit for Tat a sucker. When Naive Prober defects, however, Tit for Tat will copy that move and gain points. Tit for Tat playing against another Tit for Tat would have better results. A third strategy introduced by Dawkins is called Remorseful Prober. This one is similar to Naive Prober, except that it takes active steps to break out of runs of alternating recrimination. This strategy also needs a longer memory because it has to remember whether a defect was spontaneous or a form of retaliation. The 'remorseful Prober' allows its opponent one free hit, without retaliating. Dawkins came to the conclusion that Tit for Tat was still better.
In the next section, Dawkins defines nice, forgiving and robust. These are all words that describe a strategy. A nice on is one that is never first to defect. An example would be Tit for Tat because it is capable of defecting, but only in retaliation. Both Naive Prober and Remorseful Prober aren't because they defect, even when not provoked. A forgiving strategy retaliates, but has a short-term memory. It is quick to overlook misdeeds. Tit for Tat is also an example of a forgiving strategy. It is an example of a forgiving strategy because it does not hold a grudge. The final word that is defined is robust. This describes a strategy that is good against a wide variety of other strategies. Tit for Tat is also robust. Tit for Tat is obviously a successful strategy, but in order to remain this way, it must do well in a climate dominated by copies of itself.
Inhabitants find themselves clustered together with individuals who resemble them. This is due to the viscosity in the population. Viscosity, meaning any tendency for individuals to continue living close to the place where they were born (218). Genetic relatives tend to be similar in other aspects, not solely by looking alike, rather they resemble each other in their tendency to play or not to play (Tit for Tat). These Tit for Tat individuals may meet each other and prosper creating even more Tit for Tat individuals. They gradually grow into larger local clusters that spread out into other areas. This is not the case for Always Defect individuals. They do not prosper; in fact, they do horribly in each other’s presence. Unlike Tit for Tat, kinship or viscosity does not aid Always Defect in the population. Always Defect does not benefit from clustering.
In order for Tit for Tat to be most successful, it must always be forgiving and never be envious. For example, in a game of poker, to be envious means you do not want the other player to win more than you; your main goal is to beat the other player. Tit for Tat individuals never actually win a game. They do not think of the other player as an “opponent” because it never defects except in retaliation. He/She would prefer to join forces with the other player in order to beat the dealer.
A Tit for Tat individual also must never know the present game is the last one between them. Otherwise the chances of the players to DEFECT greatly increase. The longer the player predicts to play, the more likely he is to be nice, forgiving, and un-envious.
The threat of punishment for defection must always be present in order for Tit for Tat to be successful as well. It keeps both the players honest.
A Tit for Tat individual must be forgiving in order to make sure long runs of retaliation and defection does not occur. He/She must be willing to accept an apology as genuine.
Dawkins Ch. 12
Nice guys finish first
I. Nice guy
A. an individual that assists other members of its species and passes its genes to the next generation.
B. Nice guys can finish first
1. Robert Axelrod- reciprocal altruism
2. Prisoner's Dilemma: gambling game
a. it's a dilemma because you don't know which card to choose
b. the two choices are COOPERATE and DEFECT
c. DEFECT is always better
d. 'prisoner' term comes from imaginary example (prison sentences)
3. Prisoner metaphor
a. opponent cannot do better than DEFECT
b. one player doesn't know what the other player will play
c. there is no way of insuring trust
d. chances of getting out of jail are higher if both DEFECT
II. 'Iterated' or 'Repeated' Prisoner's Dilemma
A. similar to Prisoner's Dilemma except it is repeated an infinite number of times with the same players
1. successive rounds gives the players and opportunity to trust or mistrust
2. the players can either win a lot of money (COOPERATE/COOPERATE) or loose (DEFECT/DEFECT)
3. worst hand would be COOPERATE/DEFECT
4. example of the birds removing each other's ticks
5. strategies are limited by our ingenuity- preprogrammed rules for action
III. The winning strategy
A. Tit for Tat: professor Anatol Rapoport
1. begins by cooperating on the first move and then copying the previous move of opponent
2. what ever happens depends on the other player
3. if first player cooperates, then the next one will too and they both will continue until the end
B. Naïve Prober: basically identical to Tit for Tat except randomly throws a DEFECT and claims a high score
1. Tit for Tat vs. Naïve Prober
a. both COOPERATE for a while then Naïve Prober DEFECTS
b. Tit for Tat played COOPERATE (sucker) and Naïve Prober gains points
c. Tit for Tat retaliates- Naïve Prober copies previous move (cooperate)
d. Tit for Tat gets points- Naïve Prober is the sucker
e. Results are worse then Tit for Tat playing against another Tit for Tat
f. Naïve Prober vs. Naïve Prober do worse because DEFECTS start sooner
C. Remorseful Prober: similar to Naive Prober except that it takes active steps to
breakout of runs of alternating recrimination
1. needs a longer memory than Tit for Tat and Naive Prober
2. remembers whether DEFECT was spontaneous or retaliation
3. 'remorsefully' allows its opponent 'one free hit' without
retaliating
4. Remorseful Prober vs. Tit for Tat: runs would-be mutual
retaliation is promptly scotched
a. most of the game is spent in mutual cooperation
b. Tit for Tat vs. Tit for Tat is still better
D. Strategies
1. nice- the one that is never first to DEFECT
ie. Tit for Tat
2. forgiving- retaliates, but has a short-term memory
ie. Tit for Tat
a. grudgers don't forget- do worse because they can't break out of runs
of mutual recrimination even though opponent is 'remorseful'
b. Tit for two Tat's: allows opponents two defections in a row before it
retaliates
3. robust- a strategy that is good against a wide variety of other strategies
ie. Tit for Tat
a. robustness matters to evolutionists
b. winnings are not paid out as money, but rather as offspring
c. for a strategy to remain successful, it must do well in a climate
dominated by copies of itself
4. Suspicious Tit for Tat: technically nasty, but not very nasty- DEFECTS on
the very first move (Boyd and Loberbaum)
a. doesn't prosper because of its initial defection triggers an unbroken
run of mutual recrimination
b. Suspicious Tit for Tat vs. Tit for two Tat's: both end the game with
at least the 'benchmark', all C, score
c. Suspicious Tit for Tat scores a bonus for its initial defection
d. Boyd and Loberbaum showed that a population of Tit for Tat could
be invaded by a mixture of Tit for two Tat's and Suspicious Tit for
Tat
IV. How might mutually resembling individuals find themselves clustered together, in local aggregations?
A. Nature
1. kinship
2. animals of most species live near their sister, brothers, cousins
3. due to viscosity
a. viscosity means any tendency for individuals to continue living close to
the place where they were born
B. Tit for Tat
1. the nice strategy that is never the first to defect and it has a short
memory for past wrong-doings
C. Always Defect Strategy
V. Alexlrod’s Tit for Tat Strategy
A. This strategy is never envious meaning you’re happy as long as you and the other player are both benefiting
B. Never wins a game
C. Today, most people in a game of poker would rather beat their opponent
rather than joining together to beat the banker
1. according to Axelrod’s work, this is a mistake
D. Games
1. zero sum games: there is one winner and one loser
a. example, football
2. nonzero sum games: both players can join together to beat the
bank
a. example, Prisoner’s Dilemma
E. Individuals can benefit from one another’s success
1. cooperation and mutual assistance can flourish in a selfish world
F. Rules for Tit for Tat
1. This only works as long as both players don’t think that the present
game is their last
a. increases the probability of someone defecting
2. They should also project a long future or duration
a. increases the probability for Tit for Tat to occur
3. Players must be punished for defection
4. Players must be forgiving
5. Predictability is important to keep a pattern of mistrust
VI. Tit for Tat in Nature
A. Fig trees and Fig wasps
B. The sea bass
1. similar to Prisoner’s Dilemma
C. Vampire bats
Critical Review:
1. We thought that the results were very interesting. We were both under the initial impression that someone who is too nice will be taken advantage of and therefore not be able to succeed. Dawkins showed that it is better to be nice because you never know what the other person will do. It pays off to be nice. "Nice" does not mean to retaliate, but to only do so when necessary, then not hold a grudge. Dawkins did a series of experiments that tested which strategy would be more successful, and he came to the conclusion that Tit for Tat is the best.
2. We would have liked the author to explain the "benchmark" a little more. He mentioned it a couple times, but does not name a specific benchmark for each individual situation. We understood it t be like the goal line, but there is no specific goal mentioned.
3. In general, this chapter was poorly written. Towards the end of the chapter especially, he tried to back up his arguments with examples that were completely irrelevant or too random that could be interpreted in a variety of ways. He often lost site of his focus throughout the chapter.