Prisoners Dilema

[Caleb]

I don't know if this is the right place to post this. But, I was wondering how you would win the game "Prisoners Dilema". This is how you play it:

 

You and another person commited a crime. You can either decide to tell on the other person or stay silent. If you both stay silent, then you both serve 1 year in jail. If you decide to tell on the other person and the other person stays silent, then the person who was silent will go to jail for 10 years and you get to go free, vice versa. However if you both tell on each other then you both go to jail for 5 years.

Assuming you need to get the least amount of jail time then your partner by X amount of minutes/hours/ect. How could you win this?

[Dak]

Assuming you need to get the least amount of jail time then your partner by X amount of minutes/hours/ect.

 

eh?

 

I'd not grass: then we both go free after a year.

 

If he grasses and I don't, and I'm a criminal bad-ass enough to warrant 10 years, then I suspect that my mates will kill him; hence, he won't grass me, so I don't have to worry about that.

 

Also, I'd not grass because otherwize his mates would probably kill me

[insane_alien]

this is covered pretty well in game theory.

[Mr Skeptic]

You win by changing the scoring payouts of the game, namely by adding death to the "you go free if you tell" part.

[Greippi]

I'd not commit the crime in the first place.

 

If it's too late, I'd run around faster than the speed of light and go back in time.

[Caleb]

You win by changing the scoring payouts of the game, namely by adding death to the "you go free if you tell" part.

 

The point of the game is to have it: If you tell you may be able to get a moderate jail time or no jail time. If you don't tell then you either get little jail time or a lot.

 

Here is a game that is like this:

---

Merged post follows:

Consecutive posts merged

I've heard that in some scicology college classes, they split the class up into groups and have them play this game. If you team wins then everyone in your team gets the highest grades. If you split it then you all get roughly the same grade.

[Severian]

If I committed the crime then I deserve to be incarcerated, so going free without punishment is the worst possible outcome. Therefore, to minimise the chance of the worst possible outcome happening (to me), I wouldn't tell.

[ecoli]

the optimal strategy in game theory is to hedge your bets and go for the middle of the road. Changing strategies (mathematically) will more likely lead to the worst outcome (unless you have incentives in place to prevent it)

[Genecks]

I agree with E. Coli in a way.

 

Perhaps the best way to win in such a game is to have a backup plan that causes probability to go in your favor. If my partner in crime knows he's going to be whacked or lose some limbs if he reports me, then it's probable that he will "not confess" or choose the action that causes both of us (or at least me) to serve less time. Likewise, if my partner shares the same activity (saying I'll get whacked if I confess), then it's probable that I will not confess.

 

So, an external factor outside of the game, indirectly playing into the game, will cause probabilities to change.

 

Nonetheless, this is covered in "game theory."

[Edtharan]

Another good Game is the "Ultimatum Game" ( http://en.wikipedia.org/wiki/Ultimatum_game ).

 

I find this interesting in that the Nash Equilibriums are not what turns out to be the best solution if the game is played repeatedly with communicating agents (ie: they can tell the others what you chose to do in your last game you played).

 

It turns out that when played like this, a more fair approach to the game nets a bigger win. This is because your reputation and the willingness of individuals to make small sacrifices for the group (rejecting unfair offers), even if the group's winnings does not reflect their own winnings, means that others will be more likely to offer fair deals to them or else they risk being rejected.