human randomness game
Jul. 6th, 2005 10:24 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
guslYou know those Memory Championships?
I'd like to make a Randomness Championship, where people produce a long string of numbers/letters. People are really bad at producing random things, and tend to reuse the same kinds of patterns. Maybe it's because we can't behave as if we were memoryless.
Your performance would be measured by how much can be compressed. If your string can be compressed to less than 5% of the original size, then you've done a bad job. If you achieve higher than 90% incompressibility, that's probably better than most people.
The problem is that people could easily memorize a long random string, and so this would become really just another memorization test... unless you demand that they produce really, really long strings. Can you think of any other way to prevent this sort of cheating, besides brain-scans that would accuse any long-term memory retrieval?
I know I've already played a very simple binary guessing game, where the computer was "truly random" while learning my patterns. Obviously a game you can't expect to win.
I'd like to make a Randomness Championship, where people produce a long string of numbers/letters. People are really bad at producing random things, and tend to reuse the same kinds of patterns. Maybe it's because we can't behave as if we were memoryless.
Your performance would be measured by how much can be compressed. If your string can be compressed to less than 5% of the original size, then you've done a bad job. If you achieve higher than 90% incompressibility, that's probably better than most people.
The problem is that people could easily memorize a long random string, and so this would become really just another memorization test... unless you demand that they produce really, really long strings. Can you think of any other way to prevent this sort of cheating, besides brain-scans that would accuse any long-term memory retrieval?
I know I've already played a very simple binary guessing game, where the computer was "truly random" while learning my patterns. Obviously a game you can't expect to win.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
(no subject)
Date: 2005-07-07 07:29 am (UTC)What I'm talking about is called "Kolmogorov Complexity".
If I construct a random string of 10 digits by picking a random number 10 times
This assumption is incompatible with the notion of Kolmogorov Complexity.
(no subject)
Date: 2005-07-07 08:04 am (UTC)Right, I get the general idea.
Still, is compressibility a good measure of performance in this case? Is it not possible to construct a short program that generates an arbitrarily long semi-random string which is hard to compress using a generic compression algorithm?
(no subject)
Date: 2005-07-07 08:16 am (UTC)Yes. This was
But the seed has to be big enough to allow you to reuse it enough times without repeating a previous string.