From: Anders Sandberg (nv91-asa@nada.kth.se)
Date: Fri Dec 20 1996 - 10:58:59 MST
On Tue, 17 Dec 1996, Sean Hastings wrote:
> Anders Sandberg wrote:
> >
> > Remember that a population of TIT FOR TAT is not stable in the long run if
> > mutations and evolution occurs, defectors appear and gain great temporary
> > benefits of the overly kind population, which then collapses. And after a
> > while nice strategies emerge, and the cycle begins anew...
>
> Can you point me to your source for this, Anders?
I'll look for my references, but I don't have them handy. I have read
similar results in several places when looking at Alife preprint
archives, you might want to look there.
> I thought that TIT FOR
> TAT held its own in pools with more hostile, and nicer strategies, even
> where no memory was allowed between iterations.
Well, TFT requires 1 bit of memory.
The problem is that if you allow evolution and a somewhat gentle fitness
function, then totally nice strategies emerge in the TFT population.
Since everyone is nice, this doesn't change anything until an efficient
defector appears that can gorge on the totally nice strategies and fool
each TFT once; in these scenarios it leads to a partial or total
breakdown.
-----------------------------------------------------------------------
Anders Sandberg Towards Ascension!
nv91-asa@nada.kth.se http://www.nada.kth.se/~nv91-asa/main.html
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:35:54 MST