[ Gmail did horrible things to the original post by giving it a base64
content transfer encoding, so majordomo@vger dropped it. It's just
an off-topic digression, but I cared enough to resend anyway, fwiw. ]
On 6/14/07, Jörn Engel <joern@logfs.org> wrote:
Tit-for-tat is the best *deterministic* strategy when playing
iterated prisoner's dilemma. But note that "deterministic" and
"rational" are not adjectives that go well with "humans", and
most real-world (social) situations are noisy environments --
miscommunication and misunderstandings are the usual noise.
A double-D noise perceived by any player would throw a
tit-for-tat-playing couple into a perennial spiral of D's, for example,
which is clearly not a Pareto-efficient solution for either.
Yes, and no.
Yes - for teaching game theory (and its social relevance) in schools;
and add behavioral economics to this list :-)
No - it doesn't shorten discussions, however. And it shouldn't either.
Human / social situations are complex, Jörn; tit-for-tat can win
computer contests, for example, but it's not a behaviour one person
would find as entirely agreeable in another.
Satyam
-
