Results

[Alunyel]

I know how to really test it.

I could start a Warhammer 40,000 Dark Eldar army, and paint crosses on all of their helmets. If anyone asks, they gave up worshipping Slaanesh, and now worship God. If you all pray hard enough, and they win, then it can be nothing short of divine intervention.

Seriously, the Dark Eldar suck.

 

[dewaddict84]

program that randomly generates a sequence

AH, but that's actually kind of hard. Impossible even within the confines of pure software. You would need some sort of physical piece of hardware to collect entropy, otherwise you're just getting pseudo-random numbers which APPEAR random, but are actually sequential and pre-determined based on your initial seed number.
And that kind of hardware is very specialized, AND AWESOME:
YouTube - Dice-O-Matic

 

[Wedjat]

Wow, that's intense.

 

[Penumbra]

AH, but that's actually kind of hard. Impossible even within the confines of pure software. You would need some sort of physical piece of hardware to collect entropy, otherwise you're just getting pseudo-random numbers which APPEAR random, but are actually sequential and pre-determined based on your initial seed number.
And that kind of hardware is very specialized, AND AWESOME:

It is true that software can only generate pseudo-random numbers, but the deviation from being purely random is virtually negligible.

This summer I wrote numerous programs to test the stability of different processes, and used software random numbers a few times. I brought up the concerns to my supervisor (who is a programmer) about it not being purely random, but he said it's so negligible that for our purposes it shouldn't matter, especially since the system I am analyzing uses software random numbers anyway. I did numerous tests on generating random numbers to check how random it is.

For instance, with a simulation of 5 million numbers generated, I graphed them and recorded them extensively, and ran the simulation multiple times to see how each run compares to all previous runs. Even with a number as large as 5 million numbers generated, each run had significant variations, and overall it displayed results one would expect from random numbers. For instance, with 1 million coin flips, sometimes there would be 500,987 heads, and sometimes there would be only 498,654 heads.

One thing you can do is take the random number that comes from the seed, and put it through more processes to increase it's randomness. Like, have a computer program with a loop of 5 million generated random numbers, and for each run through the loop add the newly generated random number to the last one, so you're building a larger and larger random number that is constructed from the sum of all random numbers generated so far. Then, for each run through the loop, determine the last digit that this number ends with. If it's 0,2,4,6, or 8, then call it "heads". If it's 1,3,5,7, or 9, then call it "tails". I'm sure someone who actually has a lot of programming/advanced math experience (unlike me, with only basic programming experience and undergraduate level mathematics) could think of various mathematical ways to increase the randomness way more. There are tons of ideas online I'm sure, and don't some people do this sort of thing for doctoral thesis and such?

For the purposes of the test, if a deity can influence the laws of physics to form a certain outcome of a physical coin flip, he can similarly influence computerized coin flips.

YouTube - Dice-O-Matic

I wonder how much that cost to make...

 

[dewaddict84]

Hi. I'm a programmer. Or a software engineer if you want to get fancy.

I'd like to hammer home the point that pseudo-random numbers are not the same as actual random number. If you start up a process with absolutely no outside input, it will produce the EXACT same output each time. The seed number is the number you feed to a random number generator at the very start. It mixes splices xors and twists that number into a infinate line of pseudo-random numbers. But if you feed it the same seed, it will produce the exact same numbers each time.

It usually doesn't take much to get some sort of exterior source of entropy (read that as randomness) and a lot of new programmers simply take the current second. But that will only get you 60 variations. In Urban Dead, you used to be able to submit your actions at 8 seconds after the top of the minute and ALWAYS score a hit. Some programmers take the seconds since the epoch (Jan 1st, 1970 00:00:01), and that's better. But that means I can run some calculations and predict the outcome based on when I run your process. So if I knew the time-frame of when you generated your RSA key, even if it was a span of a few thousand seconds, I could reproduce the process in which you created the key and run those keys against the target encrypted data I'm trying to crack open. Which is why most key generators demand some sort of entropy source, like you moving the mouse around or something.

For the purposes of the test, if a deity can influence the laws of physics to form a certain outcome of a physical coin flip, he can similarly influence computerized coin flips.

No it can't. Because I KNOW what the outcome is going to be if I know the seed. That's the point I'm trying to make. The hand of god would have to do his tinkering before that seed number is entered, because after that everything is set in stone.

"The creation of random number is too important to be left to chance." - Robert R. Coveyou

(now, if you were to take the digits of Pi (3.14159...) and get a clear message out of them, I'd say that's a good argument for divinity)

 

[azmurath]

(now, if you were to take the digits of Pi (3.14159...) and get a clear message out of them, I'd say that's a good argument for divinity)

The problem with Pi is that it is infinite, which means eventually you are going to find a string with meaning to us.

 

[AV1611VET]

Pi = Be sure to drink your Ovaltine.

 

[dewaddict84]

The problem with Pi is that it is infinite, which means eventually you are going to find a string with meaning to us.

It probably does have an infinite number of decimals, but I'm pretty sure that's it not assured, yet. It could just have a ridiculous level of precision. It's been proven that it's irrational, so it can't be written as a ratio of two integers. And it's likewise been proven that it transcendental, so "there is no polynomial with rational coefficients of which pi is a root". But I'm not sure that means infinite decimals. I don't have the mathematical chops for that.

 

[Gracchus]

It probably does have an infinite number of decimals, but I'm pretty sure that's it not assured, yet. It could just have a ridiculous level of precision. It's been proven that it's irrational, so it can't be written as a ratio of two integers. And it's likewise been proven that it transcendental, so "there is no polynomial with rational coefficients of which pi is a root". But I'm not sure that means infinite decimals. I don't have the mathematical chops for that.

Actually, the fact that pi has been proven to be irrational means that it has an infinite number of decimal, hexadecimal, or binary digits.

If it had any finite number of places, however long, or if it repeated itself exactly, no matter how long the period, it would be a rational number.

 

[dewaddict84]

FACEPALM!

This is what I get for not looking through wikipedia closely enough.

Sorry.

 

[Wiccan_Child]

The problem with Pi is that it is infinite, which means eventually you are going to find a string with meaning to us.

Not necessarily. It could have every single possible combination of digits, except those with the digit '2' in them (I know this isn't true, but still). If memory serves, the property is 'transcendence'. Or you could have an infinite string of '1's, and have a '0' after every one, two, three... planes (i.e., 1.01101110111101111101111110...).

Just saying .

 

[dewaddict84]

Sorry Wiccan, I didn't quite follow that, but azmurath is correct. Since it's infinite, you WILL find ANY combination of number eventually. From the complete work of Shakespeare to the the king james bible, it's in there with perfect grammar and spelling. Now, here in the real world, finding that string of numbers might take you so long that you'll hit the heat death of the universe before you can calculate pi out to that many digits.

But that's the lesson of chaos theory. Given infinity, if it's possible, it will happen regardless of what it is.

 

[Wiccan_Child]

Sorry Wiccan, I didn't quite follow that, but azmurath is correct. Since it's infinite, you WILL find ANY combination of number eventually. From the complete work of Shakespeare to the the king james bible, it's in there with perfect grammar and spelling. Now, here in the real world, finding that string of numbers might take you so long that you'll hit the heat death of the universe before you can calculate pi out to that many digits.

But that's the lesson of chaos theory. Given infinity, if it's possible, it will happen regardless of what it is.

Only if ever possible combination of characters has a non-zero possibility of occurring in the string of digits we're looking at. This isn't necessarily the case, even if the string is infinitely long and non-repeating.

It's easy to describe such a string. For example, consider a infinitely long string of characters made up of letters in the English language, except for those with the letter 'z'. The string could be: 'a-b-c-d-e-f-g-h-i-j-k-l-m-n-o-p-q-r-s-t-u-v-w-x-y-aa-ab-ac-ad-ae-af-ag-ah-ai-aj-...' (hyphens added to clarify the pattern). Such a string would never repeat itself, but neither does it have every single possible combination of letters, since we've arbitrarily chosen the string such that the letter 'z' never occurs. You will never find the word 'zombie' in that infinitely-long, non-repeating string.

If the odds of our search word (be it zombie, the KJV, the entire works of Shakespeare, or whatever), if the odds of that word is zero, then, no matter how long the string is, you will never find it. An obscenely small chance? Sure, it will crop up all over the place. But a zero percent change? No.

 

[dewaddict84]

Ah, correct.

But I guess I have to ask why pi wouldn't contain the groups of numbers to get all the ascii characters. I read the Wikipedia entry on pi's transcendence, but uh, do you got the cliff notes version?

 

[Wiccan_Child]

Ah, correct.

But I guess I have to ask why pi wouldn't contain the groups of numbers to get all the ascii characters. I read the Wikipedia entry on pi's transcendence, but uh, do you got the cliff notes version?

I read something about pi, but I can't remember if it said you could or you couldn't find all the digits. My guess is you can. Pi is like a bad prostitute, in a way: it gives you test results you don't want, and it turns up at the most inconvenient of times.

 

[JediMobius]

I suspect the experiment failed because there was too much control. For example, not allowing the coin to bounce off anything changes possible results, and so does disallowing the flip of the caught coin onto the back of your hand. This interesting in that the result would still be random. Anyway, it's not that prayer doesn't work that way, faith doesn't work that way. You don't have to be a believer to pray earnestly, you just have to take a step of faith. Also, in my experience, although God is not adverse to showing himself through the mundane, I can't think of an example of God showing himself through the inanimate. God is a living God after all, not a coin flip.

For the sake of the hypothetical, I would have done this experiment differently. I would have found some measure of the frequency each letter is used in common english as well as spaces and used that data to decide how many r, s, t, l, n, e's I used for the code.

 

[AV1611VET]

The message according to my chart is:

Tpwoecavlyizaaxhcnemigynxaflemtiuz sorujg hh ubpjdgalbqoagaycbdmsrnuykpjzkwobefluagngepzjtzpbecaiu

Run through an anagram solver with the loosest possible parameters, it yielded zero results.

I found one!

TpwoecavlyizaaxhcnemiGynxaflEmTiuz sorujg hh ubpjdgAlbqoagaycbdmsrnuykpJzkwObefluagngepzjtzpBecaiu

 

[Bungle_Bear]

I found one!

TpwoecavlyizaaxhcnemiGynxaflEmTiuz sorujg hh ubpjdgAlbqoagaycbdmsrnuykpJzkwObefluagngepzjtzpBecaiu

13 years. Can you find "thread necromancy" in there, too?

 

[AV1611VET]

13 years. Can you find "thread necromancy" in there, too?

But one can flip a coin 800 times anytime -- even tomorrow.