game theory and artificial intelligence in mutual assured destruction

[rampart]

So, let's discuss the game with the higher stakes.
There are two players - Anton and Brad. The bet is Anton's sister Camila, kept by Brad's gang in captivity.
Possible outcomes - if Anton wins (Brad played chicken), he took his sister and they both safely leave the gameboard and Brad just lost some respect in his gang. If Anton play chiken - the gang will chop off his hands, rape his sister and then, likely, kill them both. (and Brad's rating in the gang will be increased). If Anton refuse to play - the gang will definitely rape and kill his sister and, likely, then will find and kill Anton.

What is the sane strategy in the game with those stakes?

hostage situation. finite odds that sister is harmed to some extent (from brutal rape or killed down to some less egregious indignity) no matter the play by anton.

hostage rescue is high priority but not essential. most important is deterrence of future hostage taking .

"capitulation," in the service of hostage rescue is seldom final let him think he has a way out as long as he does not escalate.

 

[Zavulon]

hostage situation. finite odds that sister is harmed to some extent (from brutal rape or killed down to some less egregious indignity) no matter the play by anton.

hostage rescue is high priority but not essential. most important is deterrence of future hostage taking .

"capitulation," in the service of hostage rescue is seldom final let him think he has a way out as long as he does not escalate.

We are not discussing the strategies of hostage rescuing. We are discussing the math question, the theory of games.

Lets just represent it as a 3D matrix.
Red are choices and possibly achieved points for Anton, blue - choices and possibly achieved points by Brad. What strategies can choose the players to maximize their prizes?

 

[rampart]

We are not discussing the strategies of hostage rescuing. We are discussing the math question, the theory of games.

Lets just represent it as a 3D matrix.
Red are choices and possibly achieved points for Anton, blue - choices and possibly achieved points by Brad. What strategies can choose the players to maximize their prizes?
View attachment 962908

just got back from a gig, and may need a few moments in the sultry southern summer evening.

it has been a few decades since i knew anything and i need to do a little refresher for the article.

also the several fictional directions that writers from azimov to maybe zelansky have taken the dream of this technology seem to have dystopian models. i don't think that is necessarily true, but WE will probably not be the ones to adapt.

 

[Zavulon]

just got back from a gig, and may need a few moments in the sultry southern summer evening.

it has been a few decades since i knew anything and i need to do a little refresher for the article.

also the several fictional directions that writers from azimov to maybe zelansky have taken the dream of this technology seem to have dystopian models. i don't think that is necessarily true, but WE will probably not be the ones to adapt.

One don't need a calculator to find solutions for the matrix that simple.
For Anton the only rational solution is "not-chicken".
For Brad, as he can read the matrix with possible outcomes and suppose that Anton's only possible strategy is "non-chicken", have to pick up "chicken" strategy.
So, Anton's prize will be 10 points, and Brad's prize will be 1 point. (Anton and Camila are saved, Brad is humiliated, but also alive. Happy end.)
Of course, even in the "chicken game", the possible prizes and losses may depends not only on the chicken/non-chicken choice, but, say, on "at what distance you turned your car".

 

[Zavulon]

Say nothing, that in real life the Mutual Assured Destruction is simply impossible, even if it were wished.