Author |
Topic: Prisoners Dilemma !! Wrong Prediction (Read 1301 times) |
|
howard roark
Full Member
Posts: 241
|
 |
Prisoners Dilemma !! Wrong Prediction
« on: Jan 30th, 2009, 6:48pm » |
 Quote  Modify
|
Studies have shown that 40% of the players actually co-operate in Prisoners Dilemma?
Any ideas on why game theoretic prediction is wrong?
|
|
 IP Logged |
|
|
|
towr
wu::riddles Moderator
Uberpuzzler
Some people are average, some are just mean.
Gender: 
Posts: 13730
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #1 on: Jan 31st, 2009, 1:11am » |
Quote Modify
|
on Jan 30th, 2009, 6:48pm, howard roark wrote:Studies have shown that 40% of the players actually co-operate in Prisoners Dilemma?
Any ideas on why game theoretic prediction is wrong? |
|
It isn't.
Game theory states what the rational choice to make is. People on the other hand do whatever irrational thing that comes to their puny minds.
Game theory =/= psychology. It has hardly any predictive value for humanity, because that is not its subject. And we're better off for it, not acting like abstract rational agents.
Evolution has primed us to act more according to what is beneficial for the species than for the individual. Game theory is therefor looking at the wrong level of abstraction.
We solve the prisoners dilemma by being irrational. Which leads to a better outcome overall (i.e. for groups, communities and the species as a whole).
|
| « Last Edit: Jan 31st, 2009, 1:16am by towr » |
IP Logged |
Wikipedia, Google, Mathworld, Integer sequence DB
|
|
|
howard roark
Full Member
Posts: 241
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #2 on: Jan 31st, 2009, 1:01pm » |
Quote Modify
|
Quote:| Game theory states what the rational choice to make is. People on the other hand do whatever irrational thing that comes to their puny minds. |
|
Quote: Evolution has primed us to act more according to what is beneficial for the species than for the individual. Game theory is therefor looking at the wrong level of abstraction.
We solve the prisoners dilemma by being irrational. Which leads to a better outcome overall (i.e. for groups, communities and the species as a whole). |
|
I was thinking along the same lines for one of the reasons.
Also peole dont follow " Game theory's Predictions" when the number of times game is played is not known in advance........or when number is infinite..........
|
|
IP Logged |
|
|
|
towr
wu::riddles Moderator
Uberpuzzler
Some people are average, some are just mean.
Gender:
Posts: 13730
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #4 on: Jan 31st, 2009, 2:21pm » |
Quote Modify
|
on Jan 31st, 2009, 1:01pm, howard roark wrote:| Also peole dont follow " Game theory's Predictions" when the number of times game is played is not known in advance........or when number is infinite.......... |
|
In the (indeterminate) iterated case the rational choice is to always cooperate. But it's not really part of people's psychology to expect things to go on indefinitely. In cases of close knit groups like family it is a fairly good prediction though (but again mostly for other reasons than rationality.)
You would probably find "public goods" games interesting too. The setting is that you take a number of players, everyone adds to the pot, which gets multiplied by some factor (say 1.5), then divided equally among all. The rational choice, again, is to defect (because you gain at no cost). But like in the prisoners dilemma everyone would gain the most if they cooperated maximally.
It becomes very interesting when you allow punishment, reputation, etc.
For example, people will, quite irrationally, pay to punish someone even if they will never play against that person again. i.e. they incur a further loss, merely for the satisfaction of wiping the smug smile off a rationalists face. However the end result is that rationalist will need to reevaluate their position and make a minimal contribution, lifting cooperation throughout a population.
But then, this can also backfire, by causing fueds where the punished punish the most cooperative players in revenge, driving cooperation down. (Behaviour which seem more prevalent in 'traditional' societies, as opposed to more liberal ones. Quite probably because they are more tuned to cooperate with family rather not strangers, as in the setting of the game.)
A link for the latter "anti-social punishment": http://research.nottingham.ac.uk/NewsReviews/newsDisplay.aspx?id=453
Another interesting game is the ultimatum game. I give you, say, $100, but that's on the condition that you split it with someone else; you decide how to split it, and the other can only veto it (in which case neither of you get anything).
Rationally, the second person should settle for whatever he can get, since it is better than nothing. But people won't settle for anything less than about 40% (on average; some people are quite generous)
You see similar behaviour of not standing for unfair treatment in fellow primates, like capuchin monkeys: http://www.livescience.com/animals/071112-monkey-treats.html
(This subject is closely related to the bachelor thesis I've been trying to write for almost a year and should have already finished long ago.)
|
|
IP Logged |
Wikipedia, Google, Mathworld, Integer sequence DB
|
|
|
Eigenray
wu::riddles Moderator
Uberpuzzler
Gender:
Posts: 1948
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #5 on: Jan 31st, 2009, 3:27pm » |
Quote Modify
|
on Jan 31st, 2009, 2:21pm, towr wrote:
Happy 10000th post!
If you found someone's wallet on the street, it would be perfectly rational to return it to the owner if the value of its contents was less than the guilt you would feel from keeping it (not to mention all the possible social rewards or costs associated with your decision). It wouldn't be the decision to return it that's irrational, but rather the decision to have a conscience, if that's really something you can control.
On the other hand, it might be rational to break a window pane and steal a loaf of bread if your sister's child was close to death, you were starving, and the probability of spending the next 19 years in jail was low enough.
Game theory cannot determine your utility function for you.
|
|
IP Logged |
|
|
|
ThudnBlunder
wu::riddles Moderator
Uberpuzzler
The dewdrop slides into the shining Sea
Gender:
Posts: 4489
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #6 on: Feb 1st, 2009, 3:47pm » |
Quote Modify
|
We have a game show on ITV that includes PD at the end.
If you Steal you are equally likely to win all or win nothing.
If you Split you are equally likely to win half or win nothing.
A no-brainer really for unscrupulous, fork-tongued contestants. (I wonder what the stats are.) Afterwards many successful Thieves predictably try to justify their duplicity by claiming to fear they would end up with nothing had they Split. For some this will be true, but if the other person is going to Steal you will get nothing whatever you do. Funny how often this supposed spoiling tactic wins all the money! I would love to be in the lift with them afterwards. Or did ITV build two well-separated lifts? LOL But it's interesting trying to predict who will Split and who will Steal based on their demeanour, body language, phrenology, etc. They have a quick chat before the denouement and obviously nobody sits there and says, "Listen love, I'll be honest with you - if I think you are going to trust me and Split I intend to Steal this money that would otherwise change you and your family's lives forever." No, "Let's do the right thing," they will coo at each other, having already decided in conspiratorial whispers with their spouses before the show to Steal, and let the Devil take the hindmost. LOL How well can you spot if someone is lying? Try to guess what these sweet-talking chancers will do.
|
| « Last Edit: Feb 3rd, 2009, 5:55pm by ThudnBlunder » |
IP Logged |
THE MEEK SHALL INHERIT THE EARTH.....................................................................er, if that's all right with the rest of you.
|
|
|
Azgard
Senior Riddler
Just gonna drive around the planet, check it out.
Posts: 453
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #7 on: Feb 1st, 2009, 4:00pm » |
Quote Modify
|
Ain't psychology grand? Especially when it comes to humans, that's when it gets really weird and interesting!
|
|
IP Logged |
Lots of opportunities are missed because they come dressed in overalls and look like work.
|
|
|
Eigenray
wu::riddles Moderator
Uberpuzzler
Gender:
Posts: 1948
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #8 on: Feb 1st, 2009, 6:29pm » |
Quote Modify
|
There's an interesting analysis on the results of a similar US show, Friend or Foe, including some demographic breakdowns and how they changed after the first season (which was completed before it started airing).
|
|
IP Logged |
|
|
|
ThudnBlunder
wu::riddles Moderator
Uberpuzzler
The dewdrop slides into the shining Sea
Gender:
Posts: 4489
|
|
Re: Prisoners Dilemma !! Wrong Prediction
« Reply #9 on: Feb 1st, 2009, 7:30pm » |
Quote Modify
|
on Feb 1st, 2009, 6:29pm, Eigenray wrote:| There's an interesting analysis on the results of a similar US show, Friend or Foe, including some demographic breakdowns and how they changed after the first season (which was completed before it started airing). |
|
Thanks for the link.
It comes as no surprise to me to notice (on Golden Balls) that younger contestants are more likely to Steal than older ones.
|
|
IP Logged |
THE MEEK SHALL INHERIT THE EARTH.....................................................................er, if that's all right with the rest of you.
|
|
|
|
RIDDLES SITE
WRITE MATH!
Home 
Help 
Search 
Members 
Login 
Register
Reply
Notify of replies
Send Topic
Print