I write a game, say Gin or Minesweeper or F-18 Attack, and I have to introduce an element of unpredictability into it to make the game playable more than once or twice. To do this, I use the built in random number function, or roll my own, which typically involves a pseudo-random progression from a seed number. For all intents and purposes, it works quite well, but it is not truly random. Or so they say. One company had to adjust the randomness on their mp3 player because users were complaining that it was not random enough, playing songs by the same artist back to back too often – perhaps it was too random.
Years ago, there was a piece in Sci-Am about Brownian Motion, the behavior of particles in a warm fluid and how one could look at paths with increasing magnification to find increasing chaoticity (w?), and it struck me that everything that happens in the universe seems to be an effect of some prior cause. If that is the case, is “random” a legitimate concept, or is it just short-hand for “so breath-takingly complex that we cannot perceive structure or trace the heirarchical causality”?
(There are sort of theological vectors to this question – the opposite of random is god – which I would prefer be placed out of the scope of discussion.)
“Any one who considers arithmetical methods of producing random digits is, of course, in a state of sin.” John von Neumann
There are several quasi-random sources on a computer. Low bits of the internal clock, mouse motion, small timing differences in keypresses, etc.
But if you want really random, you need a quantum source, which your PC isn’t going to have. Luckily, there is thing called the Internet which can supply lots of truly random bits (and I don’t mean message boards). There is the cleverly named random.org for example.
It’s a well studied problem in computer science and cryptography. See Entropy.
A program that genuinely needs randomness will generally use a strong PRNG seeded regularly from an entropy source. Single player games often use a crude PRNG; some card games take advantage of this to let you re-play the same game by providing the seed.
But is it really random? AFAIK, every quantum event still has some sort of cause, it only appears to be random because we cannot discover the causal origin (praise be to Werner Heisenberg). I do understand that particles can come into existence for no apparent reason, but can it be shown to happen without a reason?
Yes, quantum events are truly random because they meet the mathematical definitions of randomness. I can’t explain those well enough to even try but there are some people here that can and we have had several threads on it.
It isn’t true anymore that computers cannot generate true random numbers. You can buy a USB device that seeds itself based on quantum events and generates true random, not pseudo-random, numbers.
You can see more details on its official certification.
The usual term for this belief is “hidden variables”. I.e., there is some property of the particles that we don’t yet know about that “decide” when the event happens. If we knew the values of these variables, then we might predict these events better. It’s a very comforting concept to many people, including me.
Bu there are a lot of arguments against this. In addition, the ability to exploit them if they existed is pretty much nil anyway.
If there is no real way to use them to predict something, that makes it random.
One problem with hidden variables is the huge amount of information they would need to contain.
Look at an Uranium atom. It might decay now. It might decay hundreds of years from now. Or anywhere in between.
If its time of decay is fixed – if there is a “timer” – like a fuse slowly burning down – then that timer has to be able to store the decay time, and to count time between now and then, in increments of nanoseconds. That all involves a real-time clock and thousands of bits of information, including at least two “storage registers.” It requires thousands of “moving parts.”
And there is not one scintilla of evidence for any of that complex mechanism.
That is kind of ridiculous. An atomic nucleus is a very dynamic place. It is constantly flexing and vibrating, protons and neutrons circle around each other, the weak force balancing against the positive repulsion of the protons, electrons whizzing around creating sub-miniscule variations in the charge, and who knows what other particles are created and destroyed moment by moment. This is not accounting for external radiation that happens through. The “timer” is not within the nucleus itself but is the arrangement of the universe itself. How long it lasts before decaying is similar to speculating how long a Prius would last being driven hard and fast on poorly maintained forest service roads. The fact that the decay rate is more or less predictable only speaks to the fact that atomic nuclei are of relatively consistent construction. The hidden variables are there, we just cannot get at them.