random number generator/time machine

Post Reply New Topic

Ok, say you get a random number generator to make a list of random numbers, then go back in time and use the same random number generator again. Does it produce the same or different results?

This one has been bugging me for ages.

If it's a real-world random number generator, then either it relies on the internal mechanisms of a computer (in which case it isn't truly random), or it relies on data from a quantum phenomenon such as radioactive decay (which may be "truly random", depending on your understanding of physics).

The mathematics which describes quantum phenomena has probability distributions "built into it". Either you interpret these probability distributions as really being indicative of randomness ("God rolling the dice"), or you can assume that the universe is deterministic, but there is some feature of physics that we have not been able to grasp (usually called "hidden variables") which deterministically throws out outcomes that only seem random to us.

If there really are hidden variables, there are various theorems which tell us that the hidden variables are really strange. (For example, they cannot be local.) So believers in hidden variables usually believe in hidden variables because of a philosophical commitment to determinism, not because it's a nice scientific theory.

Last edited by Declension on 10 Apr 2012, 5:50 am, edited 1 time in total.

(Thread moved from Random to Computers, Math, Science, and Technology)

There is no such thing as a random number in practice (on computers anyway). They typically have an algorithm and a 'seed' value to prime the sequence. In those cases it doesn't matter what time the sequence is generated because it will always be the same. Sometimes the date/time is used as the seed for a 'random' number sequence so in that case it will be date/time dependent if it uses the current date/time.

It becomes more interesting if you use a pseudo-random number source; for example based on the radioactive decay of a sample of radioactive material using the detection or non-detection of particles emitted during a pre-set time period, building up a binary sequence as a basis for random numbers . The decay can only be analysed statistically (at the moment). However as physics progresses even this may become predictable on an individual radioactive decay basis. It would be interesting to see if a piece of radioactive material would decay in exactly the same manner if it was magically moved to another date/time. It would also depend if the decay was intrinsic to the lump of matter or if there were any external factors involved - they would obviously effect the decay if the radioactive matter was elsewhere or in a different date/time.

ok, I think i get it...
But can someone give me an extremely/moderately simple answer please, just so I know I haven't interpreted the above answers wrong. (Thanks for them btw, I didn't know how random number generators worked before hand)

OK. This is a very crude example of one way "random numbers" could be generated. Suppose you want a sequence of random numbers in the range 0 to 99. You could take pi and pick out each successive pair of digits:

3. 14 15 92 65 35 89 79 32 38 46 26 43 38 32 79 50 28 84 19 71 69 39 93 75 10 58 20 97 49 44 59 23 07 81 64 06 28 62 08 99

The distribution of values between 00 and 99 should be statistically random. However, the sequence is fixed, so it doesn't matter when you created this random number sequence, it would always be the same.

If you wanted a sequence of random numbers that was not algorithm based (like the above) and that was truly more random, you could use a balanced coin and toss it a number of times. Each time you get a head treat it as a binary 1 and each tail as a binary zero. You would end up with a random binary sequence such as this:

10011010 010111001...

If you group this into groups of 8 digits 10011010, 01011100 this corresponds to decimal values in the range 0 to 255. So you have another way of generating random numbers. Since it is extremely unlikely you would duplicate the same sequence in another time period then this could be considered a more unique and real sequence of random numbers.

"Pseudo-random" means "not random". A sequence of pseudo-random values will be being generated in a totally deterministic fashion, but using an algorithm that results in the sequence passing some tests for "randomness".

A simple such sequence comes from taking the previous 16-bit value and multiplying it by three, discarding the overflow. That works quite well for a lot of things, but using the sequence as (x,y) pairs and plotting the points doesn't look very random at all. (This was the algorithm used to generate random numbers in the Sinclair Spectrum.)

With the philosophical caveats mentioned earlier, using radioactive decay events will give a genuinely random sequence.

Sitting between those extremes, an entropy-based generator (such as Linux's /dev/random) tries to use fairly unpredictable data from the real world - the equivalent of tossing a coin.

If you travel back in time to use a pseudo-random generator, it will repeat exactly the same sequence.

So far as whether either the radioactive decay method or /dev/random would repeat the same sequence when you traveled back in time - first you have to specify exactly how you will be traveling back in time, then we can give an answer.

One of the techniques I use for random number generation is XORing bytes from compressed files (excluding file headers and footers which typically are less random). So for example using XOR between bytes in the same position in a JPEG file and a ZIP file (with different file contents of course) yields good sequences of random numbers (0 - 255). A number of years ago I also wrote a program to convert internet based radio broadcasts into wav files and from there into random binary files using a simple but effective algorithm I developed. Using such techniques you can create a method for very strong file encryption that has in effect a 100% encryption key length. The downside being of course that the encryption key is the same size as the encrypted data and still has to be transmitted from Alice to Bob somehow without Eve intercepting it.

Ok, if you had two parallel universes that hadn't deviated from each other until you press the activate button on the random number generator, would it be possible for the universes to deviate on the random number generated, assuming it was a binary\ coin flip thing. In fact, if the universes are identical in every way up to that point, is it possible for them to ever split apart at all?

The question cannot be answered in a way that can be empirically corroborated. Can you state the matter as a mathematical statement?

ruveyn

You are basically asking 'is anything truly random?' In other words, is the universe all pre-determined, or are there some events that truly do just 'pick something' without any cause?

Infinity exists on every level of everything (and infinitely doesn't exist on every level of everything), so yes/no, the universe would both split and not split. Infinity is fun.

Infinity weirds me out sometimes, but it's fun to think about.

Yes it is, assuming the many-worlds interpretation of quantum mechanics.



If the universe is fundamentally deterministic, the answer is that is should produce the same results. This really reminds of the famous "Grandfather Paradox" thought experiment. But if you apply the many-worlds intepretation, there should be a parallel universe where the results are different. I have honestly wondered if time travel, if possible, would only allow a traveler to observe events but not actually interact with the universe that they perceive in a previous time.

Observing would be interacting. It is impossible to observe without interacting and changing events. Just catching the photons of light on your retina means they don't fall elsewhere. There is a whole web of interaction and some events can cascade into very big events as per the butterfly effect. In principle something as minor as catching a photon of light could ultimately lead to the birth or non-birth of a major figure such as a government leader or major scientist.

How does on observe anything with some kind of interaction. Whatever you see is the result of photons flying into your eyes. If you are there observing you are radiating heat out into your surroundings.

ruveyn

What you're talking about is called the Observer Effect and it is a feature of the Copenhagen Interpretation. Nonetheless though, since physics depends on scale, when you scale up to the classical level these quantum effects tend to cancel out. But when you travel in time, the question remains whether any changes you make would have lasting effects when you return to the present time......And my hunch is that the answer is no because what you're seeing when you travel in time is sort of a hologram of the Universe in it's previous state.

Why dont you make the question simpler for yourself (and the rest of us).

What if you threw dice a hundred times in a row.
Or what if you spun a roulette wheel a hundred times.

And then went back in time and did the same?

Would you get the same sequence of a 100 numbers?

It would be no different than going back the next day and throwing the same dice a hundred times.

You would get a different set of 100 numbers.


But if you used a random number generator in a computer you might actually get the same sequence of 100 numbers because random number generators are not really random.