Game Theory Calls Cooperation Into Question

An anonymous reader sends this excerpt from Quanta Magazine: The physicist Freeman Dyson and the computer scientist William Press, both highly accomplished in their fields, have found a new solution to a famous, decades-old game theory scenario called the prisoner's dilemma, in which players must decide whether to cheat or cooperate with a partner. The prisoner's dilemma has long been used to help explain how cooperation might endure in nature. After all, natural selection is ruled by the survival of the fittest, so one might expect that selfish strategies benefiting the individual would be most likely to persist. But careful study of the prisoner's dilemma revealed that organisms could act entirely in their own self-interest and still create a cooperative community.

Press and Dyson's new solution to the problem, however, threw that rosy perspective into question (abstract). It suggested the best strategies were selfish ones that led to extortion, not cooperation.

[Theoretical biologist Joshua] Plotkin found the duo's math remarkable in its elegance. But the outcome troubled him. Nature includes numerous examples of cooperative behavior. For example, vampire bats donate some of their blood meal to community members that fail to find prey. Some species of birds and social insects routinely help raise another's brood. Even bacteria can cooperate, sticking to each other so that some may survive poison. If extortion reigns, what drives these and other acts of selflessness?"

Comments

Theory vs Empericism

[Wonko the Sane]

Why isn't this headline, "Game Theory Called Into Question for Failing to Predict Observed Examples of Cooperation?"

Re:Theory vs Empericism

[reve_etrange]

Why isn't this headline, "Game Theory Called Into Question for Failing to Predict Observed Examples of Cooperation?"

Unfortunately this test is all too often ignored.

Re:Theory vs Empericism

[bill_mcgonigle]

The Prisoners' Dilemma is a two-person example of a market failure - a term from economics which basically means that one actor can benefit from doing a thing that is bad for the group. Friedman's go-to example is people burning coal to heat their homes in 18th century London. Each homeowner _correctly_ calculated that coal was their personal best option, but it wasn't the best option for everybody, in aggregate. People who block intersections to avoid waiting for the next green are a more modern example - they save a little bit of time and make everything worse for everybody else.

The follow-up insight is that in natural systems, when rules result in a market failure, the rules slowly wind up changing to eliminate that scenario because those scenarios aren't good for the group (which is needed for reproduction/continuance) and (in the case of evolution) are selected against or (in the case of economics) are driven out of the economy.

Iterative plays resulting in extortion aren't going to be good for the group either and would tend to be eliminated. To be fair, we suffer broadly from the impact of psychopaths and they're only 5% of our population, so even if trends are good, local mischief isn't out of the question.

Re:Theory vs Empericism

[RazorSharp]

The phrase that makes me roll my eyes is "survival of the fittest." That's not what natural selection is. It's a gradual increase in variation with the death of the unfit. An organism doesn't have to be "the fittest," it just has to find an unoccupied niche. Thus the various "strategies" different organisms will take for survival -- be it cooperation, selfishness, or some combination of the two -- will vary depending on the organism.

Ants are pretty cooperative. Big cats are pretty selfish and territorial. But wild/feral horses are an interesting combination of the two. They have herds of mares with a few stallions. The stallions attack any other stallion that comes near and once a young stallion grows to a certain age they banish it from the herd. The stallions act pretty selfishly while the mares act rather cooperatively (however, they have a hierarchy so there's some selfishness involved, too).

I think the problem is trying to theorize a formula for understanding the behavior of organisms, or a most successful behavior, in general. There's just way too much diversity in nature for something like game theory to cover all its ground. Perhaps it works when you pigeon-hole it into capitalist economics, but I don't think it's a very comprehensive theory for explaining how animals do or ought to act.

The model isn't real.

[Kaenneth]

Real Life isn't Spherical Cows. They need a better model.

Re:The model isn't real.

[Tablizer]

It is now, Americans are overweight.

Re:The model isn't real.

[Anonymous Coward]

They need a better model.

Yes, in order to get their result they put constraints on the payoffs - and they also implicitly assume that the players can't ever stop playing.

In real life, there's typically three choices: cooperate, betray, or end the relationship. Typically as people get to know each other they increase their level of cooperation up to a level that both are comfortable with. But then if there's a betrayal the other person will often permanently end the relationship - either by walking away or by hitting the other person over the head with a large rock and hiding their corpse in some bushes.

And often there's someone else waiting eagerly to cooperate with the person who got betrayed - who will help the person who got betrayed end the relationship with the betrayer - with a rock to the head if necessary.

There are probably some situations out there in nature where their model does apply. And where cooperation doesn't happen. But it would be silly to assume that their model applied to every situation in human relationships.

Direct link to the original paper

[mc6809e]

Iterated Prisoner's Dilemma contains strategies that dominate any evolutionary opponent

The Selfish Gene

[flyhigher]

The selfish gene theory popularized by Richard Dawkins states that evolution works on genes, not on individuals. Any gene which gives rise to behavior that will cause more copies of that gene to survive, will increase its percentage in the gene pool at large.

http://en.wikipedia.org/wiki/G...

Did he read it?

[phantomfive]

[Theoretical biologist Joshua] Plotkin found the duo's math remarkable in its elegance. But the outcome troubled him. Nature includes numerous examples of cooperative behavior. For example, vampire bats donate some of their blood meal to community members that fail to find prey. Some species of birds and social insects routinely help raise another's brood. Even bacteria can cooperate, sticking to each other so that some may survive poison. If extortion reigns, what drives these and other acts of selflessness?"

I'm not sure Joshua Plotkin read the paper. It does not claim (as I understand it) to represent every scenario, merely a special case of a specific scenario. Explicitly, it requires the organism to have enough intelligence to remember what happened in previous games, so a bacteria without memory is not covered under this model. The strategy requires multiple rounds be played.

Also worth mentioning that 'good for the individual' is not the same as 'good for the species,' and nature selects the latter

I know almost nothing about vampire bats (except don't get bit, you'll need rabies shots!), but if someone understands how it relates to the prisoners' dilemma, I'd be interested in hearing it.

Nothing is possible.

[Anonymous Coward]

In other news, some people believe economic theory.

Re: Nothing is possible.

[Anonymous Coward]

It has been proven over and over that the only people who always behave according to game theory are economists and sociopaths.

Naturally we have a whole economic system based on this. People wonder why nobody's happy with it, and yet we have a population that is dumb enough to quite literally go to war for it.

Re: Nothing is possible.

[Anonymous Coward]

No. Americans are brainwashed into believing their current economic system is the only way despite the fact that it is not sustainable in the long term and is beginning to strain and fail for providing for the basic needs for nearly half the country. Then they are taught to internalize the blame for all the shortcomings of the system. If you are poor and starving it is because you are greedy and lazy, not because your formerly well paying job was outsourced to China.

Re: Nothing is possible.

[tburkhol]

Well please find an economic system that deals with the issue of sacristy, and insures its contributers exceed its detractors. At the same time insuring personal liberity.

The Kula Ring in the Trobriand Islands, where the residents of different islands developed a tradition of exchange of 'gifts,' distinct from barter-like trade. There are a number of other 'gift economies' among isolated, pre-industrial cultures. Participation is managed by social expectation and taboo, so one can argue that these systems will necessarily break down once you have enough sociopaths. One can also argue that such communities are better at recognizing and isolating sociopaths so they can't propagate their genes/behavior.

Nor is 'free market' an especially good way to deal with scarcity. If it were, then you wouldn't need social support programs. Or maybe you're going to tell me that anyone receiving social security of SNAP is not truly participating in the economy...

I don't even know where you're going with 'personal liberty.' The economic system has so little to do with what you're allowed to say, which god you're allowed to worship, or how you spend your free time as to be completely orthogonal.

Re: Nothing is possible.

[Luckyo]

What game theory has to say about that is to point out that these systems only work so long as the number of participants is small enough. Once the number of participants gets too large, it is impossible to effectively punish the leachers, and the entire system falls apart.

I guess we need to add to GP's original question the criteria of "works on a large scale"

Shachar

This is a lie often peddled in states with this system.

There are several concrete counter-examples that prove it false, ranging from Nordic countries (which view consensus and cooperation as primary tools of both political and economic systems) as well as much bigger Japan which has more of a top down system but where bosses initially even committed honourable suicide when they had to let workers go because it was considered such a significant loss of face.

These systems exist on large scale. What they require however, is a culture that promotes selflessness rather than selfishness. In the Western countries, such culture exists in Japan and Nordics. And to a lesser extent in Germany and Scotland. All of these are functional states (with exception of Scotland) where people routinely vote for and say in polls that they are willing to pay more taxes so that those who are not viable humans can live a decent life.

Re: Nothing is possible.

[duck_rifted]

So when I'm approached with offers to make more than $10k in a few weeks because I've empirically demonstrated programming skills worth that kind of pay, but I don't get the job because I don't have a degree that cost enough, it's because I'm stupid and lazy? See, the weird thing about that is, I would have thought that spending thousands of hours practicing a cognitively intensive discipline of my own volition, and becoming adept at it, meant exactly the opposite.

I honestly would never have guessed that the words "stupid and lazy" actually mean, "doesn't have parents who can afford to pay ridiculous sums for a scrap of paper, and didn't take on unnecessarily debt." And I'm descended directly from a long line of Lord Baltimores too, so it's not even a matter of pedigree. Maybe if I were smart I'd win the lottery, right?

Today's snobs are the most vacuously interesting.

Re: Nothing is possible.

[PopeRatzo]

It has been proven over and over that the only people who always behave according to game theory are economists and sociopaths.

I don't know who you are, friend, but that's the most insightful thing I've read on the internet so far today.

Of course, it's only 6:48am and I've been up for 45 minutes, but you are correct.

Re: Nothing is possible.

[Impy the Impiuos Imp]

You live in a fantasy world where those bad things are happening.

Meanwhile, in the real, actually-measured world, things are continuing to get better decade by decade.

If the west has come close to stagnating for a bit, it's because places like China are opening up and becoming more economically free. In short, the average health and wealth of economically free people continues to increase, exactly according to Julian Simon's simple, and not really controversial in the details, model.

The model doesn't describe the system.

[kuzb]

It's that simple. They have a neat mathematical model which is interesting, but if it doesn't make accurate predictions when applied to a more realistic scenario then it's missing something.

Evolution isn't about personal survival

[taustin]

It's about procreation and the survival of the genetic line. Individual survival is irrelevant, especially once one has procreated. (Though even those who don't contribute to the survival of the genetic line of their family - the person who has a sibling willing to sacrifice themselves to save the family enhances the chances the family will procreate.)

This kind of confusion is what happens when people try to do research outside of their expertise. If you want to understand biology, ask a biologist, not a physicist or a computer geek. (Though a lot of biologists make the same mistake, of course.)

Cooperation

[wonkey_monkey]

The physicist Freeman Dyson and the computer scientist William Press [..] have found a new solution to [...] the prisoner's dilemma

So... what you're saying is... these two guys have cooperated to call cooperation into question...

Riiight...

Re:Origin of *Species*

[ihtoit]

except it's the selfish assholes who equip themselves to guard what they have against anybody else getting it. The selfish assholes survive because they have the keys to the grain silo, everybody else (read: who hasn't got a copy of the key) can go die in a field. Selfish asshole (=selfish gene) survives, the meek inherit the Earth. From six feet beneath it.

Re:Origin of *Species*

[Shimbo]

You fall into the common error of equating the selfish gene with selfish individualism. Life is more complicated than that. There are times when protecting your family or the wider community at the risk of your own life is absolutely what is best for the survival of your genes.

And for sure, 'dove' is a lousy strategy in iterated prisoners dilemma. Hit him if he hits me first works fairly well though.

Re:Spite?

[bill_mcgonigle]

Spite often benefits the group over time - the individual sacrifices a little bit to increase the group benefit (which should be returned to him in aggregate as a non-zero-sum game). It can also benefit the individual by trading one thing of value for another in order to alter an adversary's value equation (like the legendary nuns who cut off their noses to avoid being raped by the marauders). Sometimes the costs are quite high but it's always a lesser-value for greater-value trade (as far as the individual values those things).

Re:Tit-for-tat is not ESS. It is well known.

[david_thornley]

The advantages of tit-for-tat are that it will stay cooperative against cooperative opponents, maximizing the total gain, and that it will not lose to its other player by more than one defection. It isn't necessarily the best strategy, but it has some provable advantages.

Therefore, if this strategy, whatever the heck it is, plays against tit-for-tat, it will come out ahead by a small amount. No extortion is possible against tit-for-tat, since it has a very short memory. Any serious attempt to hurt it hurts the opponent almost as much. The outcome of the game will be determined by the opponent, but it isn't clear to me that this is good for the opponent.

In a series of games, with players changing algorithms, tit-for-tat is not particularly susceptible to extortion, since it fundamentally yields the opponent one extra defection/cooperation win. Any attempt to extort it into more than one will fail in a competitive environment, since, if tit-for-tat is trashed the opponent is trashed almost as much. Tit-for-tat against itself, or any other strategy that won't defect first, will get straight cooperation rewards, while defect-first strategies have to accept some mutual defections, lowering the total score.

So, while I'm willing to concede that the mathematics is correct (it's been a long time since I read a mathematical paper, so I haven't checked it out fully yet), it doesn't look like it's going to make much of a difference in final score.

Re:Tit-for-tat is not ESS. It is well known.

[140Mandak262Jamuna]

The telling thing about the description, (I could not get the math) is that the strategy with longer memory loses! They are using a concept of "strategy with a theory of mind". Basically it tries to predict if the other person is rational. If they think the other one is, then it pays to be irrational. It is basically a game of chicken and the player who is reckless will win against the one who is prudent.

There is some real life applications for this. Since I am a bleeding heart liberal I see the Republicans being reckless with government shutdowns and pushing the envelop on filibusters etc as this "I will first show you I am reckless then let us play chicken". (If you are a Republican you might strongly disagree with this example).

I also see this as the explanation for being reckless revenge and disproportional response. Typically in India riots would erupt on the rumor Some boys of set A teased some girls of set B. The sets could be caste, religion, language. In the over the top response for something minor the riot inciting group suffers as much damage as all others. It is totally irrational. But the purpose is to set the stage for others for all future interactions, "Malabar Muslims or Dharavi Tamils or Biharis are known to be violent. Be more careful around them".

Or like John McEnroe's tantrums in the tennis matches is to intimidate the line judges into giving him the benefit of doubt in the future calls.

So the theory is not without its merits. But, as usual, the title is more provocative than warranted.