- #1

pianoplayer

You are using an out of date browser. It may not display this or other websites correctly.

You should upgrade or use an alternative browser.

You should upgrade or use an alternative browser.

- Thread starter pianoplayer
- Start date

- #1

pianoplayer

- #2

mathman

Science Advisor

- 7,934

- 487

- #3

matt grime

Science Advisor

Homework Helper

- 9,395

- 4

pianoplayer said:

the digits of pi are not "random" (random being a slightly ambiguous term by the way), they are *probably* normal (which is to say approcimately, that if we were to to take a string of n digits then they would occur with the correct frequency in the long run, ie about 10^-n), though we haven't proven this yet. But in general, no, there is no way to generate truly random numbers. For a start, computers tend to only operate with rational numbers (some programs symbolically deal with irrational ones) and as we know, if we restrict to the interval [0,1], then almost all numbers are irrational, ie a number picked at random from that interval is rational with probabiltiy zero, yet on a computer it will be probability 1.

However, it is possible that your son's teacher was speaking of the practical idea here, and indeed there is no way to generate truly random numbers from even any finite set, though we are able to make thigns that are "good enough" often this is some output based upon based upon the time between key strokes in the 10^{-something bigger than we can think of} time scale on you computer.

And, whislt we could in theory pick a random point in the decimal expansion of some normal number (i guess some are known), how would you decide which portion to take, randomly? and what length? and how long would it take even if we hadn't just killed the idea stone dead by asking how one picks the point we look at?

- #4

HallsofIvy

Science Advisor

Homework Helper

- 41,833

- 963

- #5

- 1,569

- 2

That's all well and good but how do you prove it when a sequence is random??

How do you prove that the best formula that generates the sequence is shorter than the list of the first n numbers in the sequence for large n?

AFAIK, you can prove that random sequences exist but that proving a given sequence is random is *hard*. In fact, the cardinality of the set of random sequences is the same as the cardinality of nonrandom ones.

- #6

EnumaElish

Science Advisor

Homework Helper

- 2,304

- 124

- #7

Nereid

Staff Emeritus

Science Advisor

Gold Member

- 3,367

- 2

What about if the PC has a hardware random number generator?

- #8

- 703

- 13

I think there is a true random number generator, that is based on quantum mechanics:

http://www.quantum.univie.ac.at/research/photonentangle/rng/ [Broken]

You let a single photon be incident on a 50-50 beamsplitter (a mirror that reflects and transmits the photon each with 50% probability). When the photon is reflected you have a 1, otherwise if the photon is transmitted you have a 0.

Then when you repeat that many times you get a sequence, e.g. 1001010110101 etc..

There are already devices in a box:

http://www.gapoptic.unige.ch/Prototypes/QRNG/ [Broken]

And here commercially available:

http://www.idquantique.com/products/quantis.htm

http://www.quantum.univie.ac.at/research/photonentangle/rng/ [Broken]

You let a single photon be incident on a 50-50 beamsplitter (a mirror that reflects and transmits the photon each with 50% probability). When the photon is reflected you have a 1, otherwise if the photon is transmitted you have a 0.

Then when you repeat that many times you get a sequence, e.g. 1001010110101 etc..

There are already devices in a box:

http://www.gapoptic.unige.ch/Prototypes/QRNG/ [Broken]

And here commercially available:

http://www.idquantique.com/products/quantis.htm

Last edited by a moderator:

- #9

- 1,569

- 2

From Wikipedia:

These processes are, in theory, completely unpredictable,...

Hmmmmmmmmmm...

I'd like to know more about this. I'm sure it has often been raised that maybe we're just too stupid to find a pattern. I'd like to see the proof that the binary sequence generated by these quantum events (like the one mentioned above) is random. Are they random in the exact same sense that flipping a coin is random or in the sense that I listed above: that the "best formula" to fit the binary sequence is not much smaller than the sequence itself?

I ask because if I made up a function, I bet I could make it so that the numbers were unpredicatble but that doesn't make the numbers a part of a random sequence.

I guess what I'm trying to wrap my head around is the math version of the statement, "these processes are completely random."

edit: Just for drill, I sought "random" in wikipedia and the first sentence is:

Randomness, should not be confused with unpredictability which is a related idea in ordinary usage.

Last edited:

Share:

- Replies
- 6

- Views
- 2K