X-Message-Number: 9736
From: Ettinger <>
Date: Mon, 18 May 1998 19:40:57 EDT
Subject: another correction

In an earlier post today I referred to one of Art Quaife's investment models,
with a predicted mean gain of 105% per year and a standard deviation of 97%. I
said this implied roughly a 5% chance of 100% loss in a year, or very roughly
a total loss once every 20 years. 

This was well off the mark, because I didn't bother--and still haven't
bothered--to make a calculation based on the actual characteristics of the
underlying distribution, which is asymmetrical and not close to the normal
distribution. 

Even so, the thrust of the observation is still valid--that an enormous
variance implies a relatively large probability of catastrophic loss, which is
not appropriate for a cryonics organization. Likely this was one reason the TT
directors were wary of making a large bet on Art's model.

Additionally, many people--myself included--are distrustful of any automated
investment or trading system. The reason is simple, viz., that markets
frequently move on PSYCHOLOGY, and we have no reliable algorithms for that. Of
course, this does not imply that mathematical models are useless--only that
they can seldom stand alone.

Robert Ettinger
Cryonics Institute
Immortalist Society
http://www.cryonics.org

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=9736