[extropy-chat] Effective relationships

[Eliezer S. Yudkowsky]

Anders Sandberg wrote:
> 
> Eliezer pointed out that there is a state one cannot get out of, being
> dead. One can of course run the whole thing with a constant inflow of new
> singles and people dropping off, but given that the death probability
> appears to be one order of magnitude smaller than the divorce probability
> these days I think it can be safely ignored.

Even ignoring death, the ergodic distribution of dyads, triads, and 
tetrads does not give you information about the relative stability of 
dyads, triads, and tetrads unless you also know the formation 
probabilities of dyads, triads, and tetrads.

E.g:  Assume that all changed relationships are from x-ad to single or 
vice versa (i.e., ignore transition probabilities between 
relationships.)  6 people have a 99% chance of forming three dyads and a 
1% chance of forming two triads.  A dyad is stable for 7 years.  A triad 
is stable for 63 years.  Ergodic ratio of dyads to triads will be on the 
loose order of 16 to 1 (I would think), but triads are 9 times as stable.

-- 
Eliezer S. Yudkowsky                          http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence