Pseudo Random Sequences

[Mark Blackman]

On 23 Jul 2007, at 09:30, Simon Wistow wrote:

> By all accounts the shuffle on your mp3 player is actually random.  
> Which
> makes sense when you think about it, a Random Number Generator (or
> Pseudo RNG for the pedants out there) is almost certainly a lot  
> chepaer
> than something that has 'favourites'. Yet despite this people insist
> that their iPods or whatever have an agenda:
>
> 	http://www.msnbc.msn.com/id/6854309/site/newsweek/
>
> Probably because the Human Brain is one massive pattern matching  
> machine
> and true randomness has, well, patterns. Randomly.
>
> But I was wondering if anybody knew of any papers out there for
> deliberately non-random sequence generators - that conciously  
> attempt to
> not have any patterns.
>
> A naive attempt would just make the probablity of a given track
> appearing be a probability function of the number of items on the list
> and the last time it was played.
>
> A cleverer attempt might have a a numbe rof different fields for each
> track and would try and make sure that it tried never play anything
> with the same value in the same field next to each other.

possibly, the obvious thing to do is treat each track like a lottery
ball and thus draw it only once until the hopper is exhausted, but would
require some way of "refilling" the hopper at some point which is  
another
UI button, so possibly sub-optimal.

- Mark

>
> I can't be the first to think of stuff liek this but my maths paper
> finding fu is weak.
>
>