Implications of Collusive Entanglement in O8

[Matt Ruff]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
You misunderstand my point. I said: If you know your opponent will chose one action (confess), then you are best off choosing that as well.

[/ QUOTE ]

But this is wrong.

[/ QUOTE ]

It's not wrong; it's factually correct. If you look at any PD matrix, if you are certain your opponent will confess then it is best for you to confess.

[/ QUOTE ]

In the immortal words of Abbott & Costello: "Third base."

I took your original statement to mean: "In the Prisoner's Dilemma, you are always better off mimicking your opponent's action." This is not true. In a non-iterated P.D., you are always better off confessing. Even in an iterated P.D., there are situations where you are best off doing the opposite of what you think your opponent will do (e.g., it’s correct to occasionally defect against a weak opponent who will not automatically retaliate).

It’s clear now that you know all this, though, so I suspect that what your original statement actually meant was: “In the Prisoner’s Dilemma, if your opponent confesses, you are better off confessing as well.” I think there still may be exceptions to this in an iterated P.D., but in general of course that's correct.


[ QUOTE ]
But again, my original point was not to elaborate on the PD, but to note how it is different than the TotC

[/ QUOTE ]

OK.


[ QUOTE ]
[ QUOTE ]
So in the case of two lows chasing the pot in 0/8, what's the finite resource?

[/ QUOTE ]

The money being added to the pot

[/ QUOTE ]

OK, that’s what I thought, and the reason I objected to this initially is that in the TotC, there aren’t supposed to be any short-term incentives for conservation, something that is clearly not true of bets into a three-way O/8 pot when you expect to get quartered.

But I think I finally see what you’re getting at: against an opponent who is determined to defect every time, and who cannot be threatened or schooled out of that behavior, the TotC suggests you are better off cooperating, while the P.D. says you should defect right back at him. So at least in that case, the TotC is a better model of the O/8 scenario, because you’re better off folding than getting into a raising war with someone who won't learn from the experience.

And by the way, “I don’t give a darn!” is our shortstop.

-- M. Ruff

[gergery]

[ QUOTE ]
But I think I finally see what you’re getting at: against an opponent who is determined to defect every time, and who cannot be threatened or schooled out of that behavior, the TotC suggests you are better off cooperating, while the P.D. says you should defect right back at him. So at least in that case, the TotC is a better model of the O/8 scenario, because you’re better off folding than getting into a raising war with someone who won't learn from the experience.

[/ QUOTE ]

Exactly, and don’t I wish I’d just typed that out at the start of this.

That’s one of my favorite comedy routines by the way.

I take it you never played shortstop either? [img]/images/graemlins/tongue.gif[/img]

--Greg