Subject: Re: is there a statistician in the house? (long)
From: "Stephen J. Turnbull" <stephen@xemacs.org>
Date: Fri, 11 Mar 2005 17:50:23 +0900

>>>>> "Marshall" == Marshall W Van Alstyne <marshall@MIT.EDU> writes:

    Marshall> 3) Gaming reputations.  The bad news is that this will
    Marshall> always be a problem.  In fact, there's a nice theorem in
    Marshall> info econ and mechanism design that states it is
    Marshall> impossible so simultaneously achieve (i) truthful
    Marshall> revelation of preferences (ii) voluntary participation
    Marshall> of all players (iii) robustness to coalitions who seek
    Marshall> to subvert the mechanism and (vi) budget balance
    Marshall> (roughly meaning that if you want the truth, you may
    Marshall> have to subsidize it in ways that generate losses).

This is basically correct, but you're missing some conditions which
put fairly strict requirements on the desirability of the outcomes
selected.  (Trivial example: the constant "do nothing" mechanism has
truthful revelation as a dominant strategy, and satisfies i-iv.)

There are alternative approaches to mechanism design which ask
"suppose we demand truthful revelation, then what outcomes can we
achieve?"  I suspect Seth is in this camp, aka "the theory of
incentive-constrained optimization".  Usually the constrained optimum
is substantially better than "do nothing."

There are also approaches which don't worry about truthful revelation,
but simply treat everything as strategic, and look at how well the
outcomes achieved satisfy the desiderata.  Again, you can usually do a
lot better than "do nothing".

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.