Random selection by computers impossible?

[marbles]

Was watching a Travel Channel show the other night on Vegas Cheats and there was one who claimed he had cracked the code of a video keno's random number generator. He made a bold and interesting statement:
(paraphrased) "It is impossible to program a computer to generate random numbers."

My wife looked at me like he was some kind of crackpot, but I agreed with him. Isn't his statement right?

[Mano]

I wouldn't say impossible, more like impracticle. You could generate numbers by doing some quantum level measurements that are truly governed by the laws of probability (like measuring the spin of neutrinos or something along those lines). Much more practical to seed a chaotic function.

[BonJoviJones]

Random Numbers To Your Desktop

[marbles]

Took about 3 seconds to get way, way over my head at that site! Thanks for the hookup, BJJ.

[BonJoviJones]

I should point out that in a sense he is right: There is no _pure software_ method of generating random numbers. Adding hardware changes that.

A major computer company (SGI?) had on their website a few years back a truly random number generator that was, mainly, a camera looking at a lava lamp.

Intel sells hardware that will generate true random numbers by measuring "thermal noise from a resistor."

I'm sure there are many other examples. There is no good excuse for not using such a system if you need secure random numbers. Of course, that didn't stop Party Poker from using using pure software for generating their "random" numbers... unthinkable...

Interesting, anyway!

[Jimbo]

BonJoviJones,

Where did you read that Party Poker uses (used) a pure software random number generator? Are you sure you are not confusing Party with Planet Poker and their trouble a few years ago?

[BonJoviJones]

Whups, I remembered it was one of the "P" pokers and knew it wasn't Paradise. My mistake, I meant Planet.

[Candyman]

I must maintain my opinion that truly random selection is not possible. Even using technology on the very edge of our understanding it is only a question of making it harder to see through the selection method. Thus, the methods mentioned make it harder (if not impossible with todays technology) to tell the difference between the generated numbers and truly random numbers, and so we might perceive them as random, but that does not mean they really are.

[Mano]

I have to disagree. Measuring quantum properties, such as the spin on a newly created neutrino, the results are truly random and not tied to some variables that we don't understand. There are equal probabilities for such a measurement to result in an up spin particle as a down spin (in fact the particle exists as both states until it is measured, at which point it instantaneously colapses to one or the other). This has been proven by physicists alot smarter than either you or I.

Hmmm, I must grudgingly back out of the discussion since my knowledge of physics is sadly lacking [img]/forums/images/icons/confused.gif[/img]. Sounds interesting though, I will look into it when I find the time.