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!gatech!swrinde!pipex!uknet!festival!dcs.ed.ac.uk!mxm
From: mxm@dcs.ed.ac.uk (Mike Moran)
Subject: Re: Minsky's new article
Message-ID: <D01LFu.D01@dcs.ed.ac.uk>
Sender: cnews@dcs.ed.ac.uk (UseNet News Admin)
Organization: Department of Computer Science, University of Edinburgh
References: <3ajo2m$ig8@net.auckland.ac.nz> <1994Nov24.235749.18364@threetek.dialix.oz.au> <3b92d9$opb@net.auckland.ac.nz>
Date: Tue, 29 Nov 1994 18:45:30 GMT
Lines: 91
Xref: glinda.oz.cs.cmu.edu comp.ai:25599 comp.robotics:15799 comp.ai.philosophy:22836

In article <3b92d9$opb@net.auckland.ac.nz> dnor01@cs.aukuni.ac.nz (David Hikaru  Norman) writes:
>telford@threetek.dialix.oz.au (Telford Tendys) writes:
>
>>> Free will is like statistics. Take a roulette wheel. On any spin, 
>>> there's a 50% chance of getting either red or black. However, if we 
>>> have complete knowledge of the situation, from the initial force 
>>> vector applied to the ball to the frictional coefficients of the 
>>> surfaces and the air, then it is feasible to predict the result. In 
>>                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>> fact, gambling cheats have done exactly this using radio links to 
>>> computers.
>
>>Pull the other one, it has bells on. Once that ball has done ten laps
>>or so, then fallen onto the knobby bits and bounced a few times there
>>is NO WAY you (or the biggest supercomputer) is going to link the
>>initial conditions to the final result. It is possible for the
>>house to cheat by slightly modifying the wheel but, although this
>>will alter the statistics at the end of the day, it won't provide
>>a prediction for a given spin.
>
>I suppose that I should be able to provide references to back up the
>things I post, but at present I can't remember the title or author of
>the book I read this in. 

	I can't remember the Author, but the title was 'The
	Newtonian Casino'.

>
>Suffice to say that it involved a keyboard built into a shoe that
>communicated with a computer somewhere else. The system worked by 
>pressing a key every time the ball went past a reference mark, from 
>which the software would figure out the likely destination of the ball 
>and provide feedback. The people involved in this were found out but not 
>prosecuted because there wasn't a specific law against this practice 
>at the time. And it worked. They made tidy sums out of it.
>
>Anyway, the point is that with a bit more information, a lot of the 
>randomness went out of the situation, with a corresponding reduction in 
>the amount of spontaneous, free-will-like behaviour attributable to the 
>gamblers.

[end of article deleted]

	It is true they did in fact make a non-trivial amount of money,
	but only on *average*. There were times when they went to the 
	casino and lost money, but as long as they could keep going
	for long enough (and avoid being detected by the casino owners)
	they would eventually start to recoup any loss.

	It is true that a large amount of the *seeming* randomness
	was removed from their perception, that is not to say that
	to an outside observe the 'randomness' would not still be
	apparent.

	The important conclusions to be drawn from this case is
	that, through chaos theory, we can see where a lot of the
	'randomness' comes from, and more importantly, we can obtain
	some sort of measure of the degree of accuracy we will have 	
	when making predictions.

	Now, in this case, we know, from the measure of the degree 
	of accuracy that we have, on the average, a chance of making
	money. This holds for this particular system, but unfortunately
	does *not* hold for all systems analysed in this way, for
	example, the weather system: the last i heard, predictions
	of weather with reasonable accuracy were still held at the
	3 weeks mark. On average any 'bet' we would make on the
	weather would probably lose us money for all reasonable
	lengths of time.

	It is for this reason that we should not hold this (Roulette) case
	up in comparison to the problem of the brains workings
	and or consciousness, and how to predict either, since	
	this example merely outlines a case where analysis through
	chaos theory shows that profitable prediction is possible. 
	This is *not* (or is unlikely to be :-) ) the case for the
	human brain.


						Thanks 

							Mike




-- 
 MI   CH AE   LM "Hes a Hypnotist | A Hypnotist of ladies |       ->   ,__o
 O R A N T H I R    You won't remember why you liked him"       -->  _-\_<,
 D  Y  E A  R  A 		    - TMBG		      ---> (*)'(*)
 I     & C     S Home: http://www.dcs.ed.ac.uk/home/mxm/
