View previous topic | View next topic

Game theory

Page 2 of 2
Goto page Previous  1, 2

scienceofsleep08
646726.  Wed Dec 16, 2009 5:47 am Reply with quote

Davini994 wrote:
I think you might need to find some new friends SoS.

Do you mean to hint that I am a misanthropic curmudgeon or a persona non grata?
At least one of those is correct :-P

P.S. A common misconception about GT on this forum seems to be this assumption that you think the other player has a 50-50 chance of choosing one of his strategies. I have no idea where people get that from...

So to elucidate the matter: In GT it is assumed that all players know the rules of the game (i.e. the other players, their strategies and payoffs) and that players are rational (they all want to maximise their individual payoff/utility). A Nash Equilibrium is then the strategy profile (set of strategies for each player) for which each individual is maximising his or her payoff, given that he knows the other players are doing the same, given that they all know he knows that they are maximising their payoff and so on...

That's the crux of it! Yes, Nash's theorem involves mixed strategies (i.e. assigning probabilities to each strategy), but let us not get ahead of ourselves and obfuscate what is essentially a simple concept.

 
Celebaelin
646731.  Wed Dec 16, 2009 5:57 am Reply with quote

scienceofsleep08 wrote:
That's the crux of it! Yes, Nash's theorem involves mixed strategies (i.e. assigning probabilities to each strategy), but let us not get ahead of ourselves and obfuscate what is essentially a simple concept.

No? Surely just a little.

 
Starfish13
646737.  Wed Dec 16, 2009 6:39 am Reply with quote

scienceofsleep08 wrote:
Starfish - IMO the concepts are analogous. The distinction is that it is widely accepted that evolution is rational, while the rationality of individual players is often disputed.

There are many examples of evolution that can be thought of as irrational, not least the 50:50 ratio of males to females in organisms which reproduce sexually (really, we don't need so much of you guys in the population). Many of these do follow through as outcomes of EGT.

Celebaelin wrote:
From an evolutionary biologists' point of view anyone who opts not to share is a 'cheater'. Social behaviour in animals has evolved because there is a genetic advantage which accrues to closely related individuals by helping one-another. Those who do not contribute in the same way as others gain advantage at no cost to themselves and, as an added bonus, to the detriment of others.

Certain conditions like eusociality in insects, use of alarm calls in populations, and female 'creches', can be explained through the idea of kin selection - the relatedness of individuals that exhibit the alturistic behaviour. However there are also benefits that occur from not cheating through co-evolution with non-kin in mutualistic symbiotic relationships, such as cleaner wrasse with their visitors, and clown fish with their anemones.

 
Celebaelin
646750.  Wed Dec 16, 2009 8:07 am Reply with quote

I agree with you but take it up with Prof. Dawkins who argues that there is no such thing as altruistic behaviour as it all amounts to the individual benefiting in some regard. I think we've touched on this before to some extent, I'll see if I can find the thread.

Nope, can't find it - but to take your example of grooming by other species both species benefit although a predator cheat, if the strategy did not eradicate the relationship, would benefit more. The example I seem to remember giving regarding the genetics of altruism was that of the selfless acts which some humans perform in helping each other, sometimes at great personal peril for no additional reward, when there is no existing genetic or direct social relationship between the persons involved. IIRC I proposed that the relatedness of humans one to another might explain this although it does not explain how an advantage is gained over other humans. Perhaps evolution has not caught up with the rapid proliferation of humans and our genetic programming leads us to believe we are all closely related but none the less if we were all to have this tendency to see others as part of the tribe then the advantage would be shared by all.

On a slightly separate point it should be mentioned that humans do not have an alpha male (or an alpha female) as such so, deliberately or otherwise, breeding diversity is an element in our species survival strategy.

 
Gray
646783.  Wed Dec 16, 2009 9:52 am Reply with quote

Dawkins (quoting someone else, IIRC) explains this phenomenon by saying that humans always used to live in small populations where everyone you knew was probably related to you, and although we now don't know our neighbours, this behaviour has remained somewhat in our brains (although not in many people, of course).

There's also the idea championed by Sue Blackmore which I think is very interesting, and follows a genetic/memetic 'public credit' co-evolution based system of explanation. Helping someone else in an especially dangerous situation will ensure that everyone around you gets to hear about the amazing rescue (or whatever) and therefore increases your chance of finding a mate who is impressed at this act of selfish heroism.

In this case, the memes and the genes both benefit (by being copied into their next generations): the genes will go into children, and the memes will go into the minds of everyone who hears about the amazing story.

And then there's the much more interesting version to do with cats, which I think is spectacular because it raises great questions about free will:

http://www.edge.org/documents/archive/edge307.html

 
Jenny
647013.  Wed Dec 16, 2009 9:30 pm Reply with quote

That is a fascinating article Gray - thanks for that link.

 
Starfish13
647038.  Thu Dec 17, 2009 5:12 am Reply with quote

Gray wrote:
Dawkins (quoting someone else, IIRC) explains this phenomenon by saying that humans always used to live in small populations where everyone you knew was probably related to you, and although we now don't know our neighbours, this behaviour has remained somewhat in our brains (although not in many people, of course).


Could this phenomenon be something that might be expected to proliferated through the human population in this era of social networking - you may not directly know friends of friends on facebook (for e.g.), but you may be in contact with them in your daily life, and word gets around (as with sue Blackmore's explaination).

 
Davini994
647110.  Thu Dec 17, 2009 9:22 am Reply with quote

SoS wrote:
Do you mean to hint that I am a misanthropic curmudgeon or a persona non grata?
At least one of those is correct :-P

No, I was observing that you seem to assume that people act like barstewards - so perhaps time to meet some new people!

The altruism seems pretty simple to me - if a species is genetically predistopsed to be altruistic then they have a competetive advantage as collaboration is massively advantageous.

Being nice to people seems to cheer the person up, from what I can tell.

Am I oversimplifying again?

 
Starfish13
647144.  Thu Dec 17, 2009 11:13 am Reply with quote

Davini994 wrote:
Being nice to people seems to cheer the person up, from what I can tell.


Be excellent to one another. Dude.

 
Celebaelin
647160.  Thu Dec 17, 2009 12:18 pm Reply with quote

Gray wrote:
In this case, the memes and the genes both benefit (by being copied into their next generations): the genes will go into children, and the memes will go into the minds of everyone who hears about the amazing story.

So if human beings collectively reward selfless heroism generally (tbh I'm not sure that they do to any great extent) it then ceases to be selfless and becomes a desirable trait? Judging just from the prevalence of this sort of behaviour in human populations I don't think the rewards match the risks. We should expect the incidence of altruistic behaviour to reflect the rewards generated if it were a purely genetic matter but you have already raised the point of memes. If we substitute the phrase 'idealised social behaviour' then this becomes a bit clearer - the altruistic action is the one most people would praise and promote in others as it has a potential benefit for them to do so. On the other hand there is less mileage in praising someone for a selfless act they have already performed - that grants reward without necessarily promoting gain for the praiser. For example the fire brigade are often eulogised generally as a body but very few people actively single out individual firemen or even say 'firemen are idiots' or 'fireman X is an idiot'; there is a suspicion though that their wages may not match the risks they take which goes some way to implying that. As for cats, toxo and recklessness that is fascinating but I doubt the behavioural influence has a great effect on the incidence of altruistic behaviour (though in truth I have no evidence to base that doubt on).

Getting back to something a bit more akin to Game Theory if you were to allocate realistic numbers and probabilities to these situations the hard facts will most likely always favour the cheat as (s)he always makes the selfish choice. This is however viewing a single instance, as starfish has said evolution works over multiple repetitions of similar situations and cheaters run the risk of being identified and suffering the consequences of being exposed. Cheating in the 'fireman' example means not paying your taxes (which reminds me...) but there are other examples of the 'good neighbour' variety which identify humans as social animals and cheaters as, well, something less than the typical stoic human quietly making things work out better for those around them without the expectation of reward - naming no names of course!



Bill & Ted wrote:
You'll see!

 
Davini994
647231.  Thu Dec 17, 2009 5:38 pm Reply with quote

... and that's why tories should be given a good kicking wherever possible.

;)

 

Page 2 of 2
Goto page Previous  1, 2

All times are GMT - 5 Hours


Display posts from previous:   

Search Search Forums

Powered by phpBB © 2001, 2002 phpBB Group