Random numbers are so important that sometimes at some point we do feel "Are the random numbers really generated randomly?" and "Can such an important topic be really left to chance?".What do you guys feel is the answer to this?

No. Care must be take in finding a quality "random" generator.

There was a case where a gambler found that the generator, of some video machine, was seeded by the current time, and the machines were turned on at the same time everyday. So he found the sequence and got a bunch of cash.
Another similar case I believe where the machines used a static string of numbers and so someone found the sequence, used it, and told others how to use it.

So that leaves us no more ahead than NULL !!! We haven't generated truely random numbers .... !!! :(

Well , I believe that , Random Numbers arent really random at all, Even if you ask a human to tell a sequence of random numbers they are again dependant on the mentality of a human being i mean his likes or dislikes , Date of Importance to him, What number he feels is lucky and so on.

So How can we say that a random number is actually random. When we have many different constraints added to it.

For Example In the Game Bingo - each time a number is removed from the total set. The Probability of the other numbers increases.

So The Definition of a Number being random might be un-defined.

So Randomness again comes back into being constrained by many boundaries , So I think that Randomness is kind of Impossible to Achieve.

commented: Great post...I totally agree... +2
commented: scru is just a random idiot :P +9
commented: apparently, this topic drives scru insane. +10
commented: oh shut up -1

there's no such thing as a truly random number. all random number generators are pseudo-random.

the difference is that some generators are better than others, in that their sequences are harder to predict than others, given current computing technology.

in other words, we have number generators that are, for all intents and purposes, "practically" random.


commented: disapprove -1
commented: blah. +11
commented: True :) +3
commented: you shut up too -1

Oh yes, the well ordered world of random numbers. The casinos couldn't live without.

Lotteries are based on RANDOM NUMBERS and it amazes me what it comes up with!

People say the the lottery IS RIGGED so you cant win,but im telling you,IT ISNT :icon_cheesygrin:

I created a program on my commodore64 years ago,a LOTTO program.. It randomizes and picks up to 10 numbers from 1 to 80 and you have you enter which #'s you want....... It amazes me how many times i see consecutive numbers come up!! (61,62,63,64,65,66 came up one time)

I DIDNT RIG ANYTHING!!!!! :icon_cheesygrin: :icon_cheesygrin:

You wouldnt ever choose a string like that (And if you did,they probably wouldnt come up :D)

The mystery of randomization!

commented: lol +10

your Commodore64 probably didnt have a very strong randomize function.

i also assume you were about 10 years old, or so. sorry if i don't spend much time analyzing your memories of code you wrote in the early 1980's.


It amazes me how many times i see consecutive numbers come up!! (61,62,63,64,65,66 came up one time)

What did you use in your generator: current time?

Nope i used standard randomize routines....

I have seen the same thing in the REAL QUICK DRAW also!! (Consecutive strings)

I still use it quite often and i cant ever get anything good :D (Just like in the real quick draw (UUGGHH))

I'm surprized at how much game code I see has weak RNG source code.

The first rule of RNG is that you need a true random stream of bits as source data. This stream must at the very least pass the FIPS 140-1 testing, Marsaglia Die-Hard test, and th Berlekamp-Massey examination and show no statistical weakness.

Once statistically valid, then that stream is utilized in the generation of numbers as well as shuffling, etc., which also has to be quantified as having no statistical weakness.

Some RNG's are dedicated systems that generate a data stream, do analysis upon that stream and if it either fails or demonstrates an anomaly, then jettisons that stream and generates another!

Some of the best RNG bit streams are from hardware generated noise and implemented in an entropy situation.

As a random adition, i did a test on PHPs rand() function to get it to generate eather a 1 or a 2 and found that
1 came out 50.4% of the time
2 came out 49.6%
Even random numbers such as balls in a bingo machiene are not realy random because if you put every ball in the exact same position with everything about the environment the same, you will get the same result.

How many times did you flip your coin?

Try it 10,000 or 100,000 times! The less sampling the more it'll be off. You will find that it is almost never 50.000%

for every random act there are millions of procedures to be done

One of the methods I know have been used is to generate a pseudo-random # then move the decimal a pseudo-random # of places to the right and go to that location of the calculation of pi (see here if you want the first 100 billion digits of π).

Most of the successful methods that I have heard of use variations of this (rng seed is generated by choosing from the list, etc). As far as I know, the next digit of pi is the most random number available; of course, once you have a list - is it random any more?

AH, "Grim Jack" I detect a pattern!

Would be interesting to run a statistical analysis on the numbers that method genertes!

One of 'interesting' searches suggested by one of my heroes (Isaac Asimov) was that, if there was a god/God, he would hide a message to us in PI.

Odd facts:
The Guinness-recognized record for remembered digits of π is 67,890 digits, held by Lu Chao, a 24-year-old graduate student from China. It took him 24 hours and 4 minutes to recite to the 67,890th decimal place of π without an error.

The Feynman point, a sequence of six 9s that appears at the 762nd through 767th decimal places of PI (Feynman wanted to memorize up through the 767th place and end with "999999 and so").

The sequence 9876543210 occurs beginning at digits 21981157633, 29832636867, 39232573648, 42140457481, and 43065796214.

The distribution of digits in 1/PI is roughly as randomly distributed as the digits of PI.

Generally, when talking about calculating PI, people are actually calculating PI - 3.


Find the prime

but I digress.