## A easy problem

A forum for good logic/math puzzles.

Moderators: jestingrabbit, Moderators General, Prelates

Amicitia
Posts: 618
Joined: Wed Aug 29, 2007 5:37 am UTC

### A easy problem

Two economics post graduates convicted for crimes against humanity are trapped in a prisoner's dilemma. What does each do in order to maximise their payouts, knowing their opponent is doing the same?
"OMIT NEEDLESS WORDS"
William Strunk, Jr.

Macbi
Posts: 941
Joined: Mon Apr 09, 2007 8:32 am UTC
Location: UKvia

### Re: A easy problem

Spoiler:
If their opponent has to do the same thing then they should stay silent.
I often think that (in the normal dilemma) as both are perfectly logical, and they know this, and their sitiuations are identical, they know that they must do the same thing.
Indigo is a lie.
Which idiot decided that websites can't go within 4cm of the edge of the screen?
There should be a null word, for the question "Is anybody there?" and to see if microphones are on.

Ari
Posts: 725
Joined: Sun Oct 07, 2007 5:09 pm UTC
Location: Wellington, New Zealand

### Re: A easy problem

A single-shot dilemma?

Well, that depends if they're rational or superrational.
"Hey %*&^er, offensive communication works fine so long as you do it respectfully."
"I am so quoting that out of context at a later date."

Robin S
Posts: 3579
Joined: Wed Jun 27, 2007 7:02 pm UTC
Location: London, UK
Contact:

### Re: A easy problem

By "knowing their opponent is doing the same", do you mean "knowing their opponent will choose the same action as them" or "knowing their opponent will follow the same lines of reasoning as them"?
This is a placeholder until I think of something more creative to put here.

Macbi
Posts: 941
Joined: Mon Apr 09, 2007 8:32 am UTC
Location: UKvia

### Re: A easy problem

Robin S wrote:By "knowing their opponent is doing the same", do you mean "knowing their opponent will choose the same action as them" or "knowing their opponent will follow the same lines of reasoning as them"?

Their situations are identical, so I think the two are equivalent.
Indigo is a lie.
Which idiot decided that websites can't go within 4cm of the edge of the screen?
There should be a null word, for the question "Is anybody there?" and to see if microphones are on.

EdgarJPublius
Official Propagandi.... Nifty Poster Guy
Posts: 3712
Joined: Tue Oct 09, 2007 4:56 am UTC
Location: where the wind takes me

### Re: A easy problem

No, that's the essence of the Prisoners Dilemma, that tho two aren't equivalent.
Roosevelt wrote:
I wrote:Does Space Teddy Roosevelt wrestle Space Bears and fight the Space Spanish-American War with his band of Space-volunteers the Space Rough Riders?

Yes.

-still unaware of the origin and meaning of his own user-title

Token
Posts: 1481
Joined: Fri Dec 01, 2006 5:07 pm UTC
Location: London

### Re: A easy problem

EdgarJPublius wrote:No, that's the essence of the Prisoners Dilemma, that tho two aren't equivalent.

Not really. If you are reasoning the same as someone, you'll pick the same answer given the same information. The essence of the Prisoner's Dilemma is that the Nash equilibrium is not the "optimal" choice.
All posts are works in progress. If I posted something within the last hour, chances are I'm still editing it.

zingmaster
Posts: 480
Joined: Wed Oct 24, 2007 6:22 pm UTC
Location: Ha!
Contact:

### Re: A easy problem

And so since they're economists assumedly working together, they know the payout matrix and have agreed beforehand to avoid Nash Equilibrium.
You get 500 xp.
You collect:
1 HOBO BONUS
1 CHAOS BONUS
1 rusty dagger

Hold on Dreamaway

Ralp
Posts: 25
Joined: Sat Dec 16, 2006 9:44 am UTC
Location: internet
Contact:

### Re: A easy problem

Token wrote:
EdgarJPublius wrote:No, that's the essence of the Prisoners Dilemma, that tho two aren't equivalent.

Not really. If you are reasoning the same as someone, you'll pick the same answer given the same information. The essence of the Prisoner's Dilemma is that the Nash equilibrium is not the "optimal" choice.

If you reason the same as your opponent, and your situations are identical, then you'll both play by the same strategy, but that doesn't mean every or even any move will be identical. We might both (independently) decide to cooperate with probability p initially, and thereafter cooperate with probability q or match our opponent's last move with probability 1-q.

Incidentally I think what I just described is a pretty good general strategy for the iterated game (for appropriately chosen p and q), but if they both know the other is a post-grad econ student (as Common Knowledge), I don't see how the optimum strategy wouldn't just be "always cooperate".

Token
Posts: 1481
Joined: Fri Dec 01, 2006 5:07 pm UTC
Location: London

### Re: A easy problem

Ralp wrote:Incidentally I think what I just described is a pretty good general strategy for the iterated game (for appropriately chosen p and q), but if they both know the other is a post-grad econ student (as Common Knowledge), I don't see how the optimum strategy wouldn't just be "always cooperate".

It's the optimum strategy no matter what each person knows about the other. However, there is a difference between an optimum strategy and an equilibrium strategy. Always cooperate is the former, always betray is the latter.
All posts are works in progress. If I posted something within the last hour, chances are I'm still editing it.

Ralp
Posts: 25
Joined: Sat Dec 16, 2006 9:44 am UTC
Location: internet
Contact:

### Re: A easy problem

Token wrote:
Ralp wrote:I don't see how the optimum strategy wouldn't just be "always cooperate".

It's the optimum strategy no matter what each person knows about the other. However, there is a difference between an optimum strategy and an equilibrium strategy. Always cooperate is the former, always betray is the latter.

It's certainly not the optimum strategy for me if I happen to know you're playing "always betray", and it might not be the optimum strategy if I know there's a good chance you will betray me sometimes (depending on that probability and the specific payoffs). I think it's also pretty important whether I know anything about whether my potential cooperation or betrayal might affect your future decisions to cooperate or betray me.

Macbi
Posts: 941
Joined: Mon Apr 09, 2007 8:32 am UTC
Location: UKvia

### Re: A easy problem

Ralp wrote:
Token wrote:
EdgarJPublius wrote:No, that's the essence of the Prisoners Dilemma, that tho two aren't equivalent.

Not really. If you are reasoning the same as someone, you'll pick the same answer given the same information. The essence of the Prisoner's Dilemma is that the Nash equilibrium is not the "optimal" choice.

If you reason the same as your opponent, and your situations are identical, then you'll both play by the same strategy, but that doesn't mean every or even any move will be identical. We might both (independently) decide to cooperate with probability p initially, and thereafter cooperate with probability q or match our opponent's last move with probability 1-q.

Incidentally I think what I just described is a pretty good general strategy for the iterated game (for appropriately chosen p and q), but if they both know the other is a post-grad econ student (as Common Knowledge), I don't see how the optimum strategy wouldn't just be "always cooperate".

They only play one game so multiple game strategies don't work (although the one you suggested is good).
Actually, looking at the first post that might be false, but if they both cooperate in the first game (which they should) then your strategy is equal to mine.
In one game: If they both cooperate with probability p then their expected gain is highest when p=1.
Indigo is a lie.
Which idiot decided that websites can't go within 4cm of the edge of the screen?
There should be a null word, for the question "Is anybody there?" and to see if microphones are on.

Indon
Posts: 4433
Joined: Thu Oct 18, 2007 5:21 pm UTC
Location: Alabama :(
Contact:

### Re: A easy problem

If I know my opponent would use the exact same reasoning as I will, thus leading to the same result, I will choose to stay silent, regardless of if my opponent betrays me... which I know he won't, because he's following the same line of reasoning.

More, uh, reasonably, though, I would play the game as I would in any other circumstance, because knowing what action someone will choose (which is what knowing their strategy essentially means) seems to trivialize the scenario.
So, I like talking. So if you want to talk about something with me, feel free to send me a PM.

My blog, now rarely updated.

morjax
Posts: 50
Joined: Mon Apr 23, 2007 4:00 pm UTC
Location: Twin Cities, Minnesota

### Re: A easy problem

Token wrote:
EdgarJPublius wrote:No, that's the essence of the Prisoners Dilemma, that tho two aren't equivalent.

Not really. If you are reasoning the same as someone, you'll pick the same answer given the same information. The essence of the Prisoner's Dilemma is that the Nash equilibrium is not the "optimal" choice.

"What does each do in order to maximise their payouts, knowing their opponent is doing the same?"

They aren't necesarily reasoning the same, They are only both trying to maximize thier payouts, knowing that their opponent is also trying to maximize her payouts. Also, How can we be sure that our opponent is reasoning soundly?

Robin S
Posts: 3579
Joined: Wed Jun 27, 2007 7:02 pm UTC
Location: London, UK
Contact:

### Re: A easy problem

I posted here about Iterated Prisoner's Dilemma strategies. Ignore the first paragraph.
This is a placeholder until I think of something more creative to put here.