faux darwinism, entropy
Mon, 18 Feb 2002 11:35:52 -0800
> Capitalism only works
> if competition allows the best to rise to the top, Darwin-style.
Where does "the top" lie, for real
biological Darwinism? Should we
go by total biomass, by number of
> So, both MS and AOL woke up one morning and had control of 90% of the desktop
> and a majority control of the IM users? They did nothing better than their
> competition to get to that point?
While I believe there are some
things that MS and AOL did well,
large marketshares may not imply
high fitness if games are close
to zero-sum. *
After enough rounds of zero-sum
games, most players will be wiped
out. The remaining players will
be very large. This is true even
for initially identical players.
Suppose I offer to play a game
with you: we flip a coin, until
you say "stop". If it always
came up heads, I owe you $2^H.
If it ever came up tails, $0.
How much would you pay to play?
How long would you let it ride?
Show me a market with a hockey
stick distribution, and I will
tend to believe that we should
rule out time and chance (or,
among Machiavellians, Fortuna)
before looking for more complex
(For the more realistic version,
say I have a coin that you know
to be unbalanced so it comes up
heads 55% of the time. How much
of your stake would you bet on
each flip? Will you change your
answer if instructed to maximize
mean return? What about modal?)
* even positive-sum games may be
driven close to zero-sum by some
In poker, one possible strategy
is to chase weaker players (who
may yet improve) out by raising
until they no longer face a good
bet. The raise lowers everyone's
expected profits, but if one has
the strongest hand, that's fine.
Are industries with huge amounts
of Turpin's Pixie Dust involved
an example of industrial poker?