X-Message-Number: 2350
From:	Ralph Merkle <>
Subject: CRYONICS Steve Harris as an Alcor Board Member
Date:	Fri, 16 Jul 1993 17:27:41 PDT

Steve Harris writes:
>As regards this last, I am of course glad to see Brian Wowk toss
>his hat into the ring as an Alcor Director candidate for 1993. 
>If it will help this situation, I myself am willing to follow his
>example and do the same.

I strongly support the election of Steve Harris to the board of
Alcor.

As many of you know, Steve was a candidate for the board last year,
but did not get elected.  A likely reason for this is the
secrecy involved in the vote, the resulting uncertainty of
how others will vote, and the simultaneous need to estimate how
others will vote to determine your own optimal voting pattern.
This results in voting patterns and instabilities in the voting
system that are much worse than the instabilities required by
the various theorems on voting that show that no voting scheme
can satisfy all of a set of basic requirements.

As an example, suppose all members of the board assume that candidate
X is wonderful and will certainly win.  Then each board member will
assume that the other board members will vote for X, hence each board
member will decide that the optimal voting strategy is to use their
scarce votes to vote for some *other* candidate (because X will surely
win, while the more marginal candidate preferred by this particular
board member might not).  Hence, all board members fail to cast their
votes for X, and X loses.

This illustrates how a candidate strongly preferred by all board
members can lose an election to the board under the current voting
system.

The simplest method of eliminating this instability (and a broad range
of other, related instabilities) is to have all board members reveal
their votes to all other board members.  Note that a public vote,
e.g., revealing votes to all members, is not necessary to eliminate
the instability described above.  The board members need only discuss
and reveal their votes to each other.  This will result in a "negotiated"
board.  The board members would retreat into private session, and would
then negotiate over the composition of the new board.  Once the negotiations
terminated, (all board members figuring they have accomplished as much
as they can accomplish using their votes), then the new board is announced.

It is theoretically possible that this method of voting will not terminate,
e.g., that two board member's voting behavior both depend on the other board
members in an unstable fashion.  This problem is indeed inherent in any voting
scheme (unlike the more severe instability described above, in which the board
fails to elect a candidate who all board members desire to have on the board).
A simple method of resolving this problem which is as satisfactory as any
other would be to (1) have the board members adopt an unbiased order of
themselves (alphabetical by last name is fine); and then (2) have the board
members each cast a single vote in the specified order, and then (3) repeat
step (2) until either all votes have been cast, or all board members have
elected not to cast their remaining votes.

I *very* strongly urge the board to eliminate secrecy among themselves during
the voting process.  Such secrecy can lead directly to bizarre voting patterns,

voting patterns that *no one* desires.  Whether or not the board would also wish

to make the vote public is a second issue, which the board might or might not 
wish
to do as it sees fit.

Ralph C. Merkle


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=2350