Newsgroups: comp.ai,comp.robotics,comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!pipex!sunsite.doc.ic.ac.uk!dcs.gla.ac.uk!unix.brighton.ac.uk!mjs14
From: mjs14@unix.brighton.ac.uk (shute)
Subject: Re: Minsky's new article
Message-ID: <1994Nov23.111502.28897@unix.brighton.ac.uk>
Organization: University of Brighton, UK
References: <3ajo2m$ig8@net.auckland.ac.nz> <3ak1n5$ndn@mp.cs.niu.edu> <3ap502$ed9@net.auckland.ac.nz>
Date: Wed, 23 Nov 1994 11:15:02 GMT
Lines: 56
Xref: glinda.oz.cs.cmu.edu comp.ai:25414 comp.robotics:15592 comp.ai.philosophy:22491

In article <3ap502$ed9@net.auckland.ac.nz> dnor01@cs.aukuni.ac.nz (David Hikaru  Norman) writes:
>OK, free will allows you to choose the best option. I infer from
>this that it also allows you to choose the worst option. However,
>if you really want to win at roulette and you know that there is a
>"best" way of doing this (and I can't see any subjectivity about it, 
>but you can try and convince me otherwise if you like), then you are 
>never going to choose anything but the best option.

Isn't the best option at roulette simply not to play in the first place?
(And thereby not incur any losses to the Banque)  :-)   :-)

>Unless your purpose is something different  [...]

Absolutely!  Nail on the head hitting thereof.

Everyone in this thread seems to keep reverting to the temptation of
assuming that decision making occurs in one-dimensional space.
In reality, decisions are made as a complex tradeoff between many many
parameters, in the hope of finding a compromise solution that just gives
us an edge in our overall quest for health and happiness.

Even Sarah Boynton (to the surprise of others around her, who thought
that they could predict her choice) might choose to turn down an offer of
chocolate on odd occasions, when other compromises and goals are taking
precidence.

>The only occasions where decision-making is required is when the best
>possible option is unknown, either because of a lack of data or an
>inability to process all the data.

I think you've reverted to thinking one-dimensionally again.
Humans who are tackling just one problem at a time... only when that is
achieved (or ditched as a bad job) does the human move on to the next.
In reality, humans are juggling many problems at a time... sometimes
thinking of solving one, sometimes thinking of solving another.

>This is where rules of thumb and
>learning by experience kick in, so as to prepare you for the next
>time that you face an analogous situation.

True.

>Casino operators make a tidy profit with such random things as
>roulette wheels, by knowing the statistics involved. This doesn't 
>sound particularly irrational or stupid to me.

Absolutely.
But what about the players?
May be they have some other goal that takes precidence over the lessons
that they learned in statistics.  Maybe they are there for something else...
the social life, for instance.

Just my thoughts.
-- 

Malcolm SHUTE.         (The AM Mollusc:   v_@_ )        Disclaimer: all
