From: David Sternlight <david@sternlight.com>
Date: Fri, 26 Jul 1996 18:28:37 +0800
To: Hal <cypherpunks@toad.com
Subject: Re: Twenty Bank Robbers -- Game theory:)
In-Reply-To: <199607252305.QAA06996@jobe.shell.portal.com>
Message-ID: <v03007801ae1e1cb77f7f@[192.187.162.15]>
MIME-Version: 1.0
Content-Type: text/plain


At 4:05 PM -0700 7/25/96, Hal wrote:
>I think the best way to approach this problem is to first try to solve
>it assuming there are only two robbers rather than 20.  Then once you
>have that figured out, try it for three, then four, and so on.  Keep in
>mind that 50% support is enough for a proposed distribution to pass, you
>don't need a strict majority.
>

Exactly. I arrived at the solution the same way. Note that there is another
assumption needed--that the selection of a proposer is by lot at each new
stage. If the ordering of proposers is known in advance, a different
solution results.

A further assumption is that a certainty gain of 1/n of the total sum is
preferred to a 1/n probability of gaining the entire sum and a (1-1/n)
probability of gaining nothing..

David