Thread: Can a human simulate a random generator?

Is it possible for a human to simulate a random number generator, just by making up numbers and trying to be 'random'?

8 5 5 1 9 2 9 9 0 9 2 2 6 4 6 1 2 5 2 9 1 9 1 1 1 7 0 1 6 2 9 8 8

That's my attempt. I tried to repeat some numbers as much as a machine would by chance. I doubt I could, but is it possible for any human to actually simulate one correctly, without falling victim to a tendency to choose certain numbers over others at certain points in the pattern?

(Random number generators themselves are only simulations of randomness, since they follow formulas, and once you plug in the system time in milliseconds as a seed in the formula, it becomes so unpredictable it might as well be random.)

I don't believe that a random number generator is possible.

Any machine is still going to be bound by the laws of physics, and the formation of the particles at the moment of the big bang (or whatever the hell you believe) and therefore is predetermined.

I also don't believe in predetermination. No matter how much one can calculate the events and therefore the formation of the particles at any given moment in time there is still the question of why anything is the way it is. Even if every particle in the universe was uniform and aligned one could argue that this fact, in itself, is very random.

So, to answer your question:
I believe that a human can generate random numbers just as well as a machine because everything is relative. Even if I were to just name off a bunch of twos, I think that in itself would be fairly random, considering the fact that I had an infinite number of numbers to choose from.

The only exception to this rule is Tom Cruise. He is completely fugging predictable and lame.

Then you don't know quantum physics.

I think the question needs more specification as to what kind of process you're allowing. Can you take cues from your environment? Can you do some kind of mental algorithm?

An interesting and surprising piece of related mathematics: Benford's law - Wikipedia, the free encyclopedia

This has tripped up quite a few data falsifiers in the past, including Iran, I think.

Originally Posted by Xei

Then you don't know quantum physics.

I'm not really impressed with the theories associated with quantum physics anymore.

Originally Posted by Xei

I think the question needs more specification as to what kind of process you're allowing. Can you take cues from your environment? Can you do some kind of mental algorithm?

An interesting and surprising piece of related mathematics: Benford's law - Wikipedia, the free encyclopedia

This has tripped up quite a few data falsifiers in the past, including Iran, I think.

The question probably wouldn't need so much specification if I hadn't complicated the crap out of it..

I'm sorry. :-\

Originally Posted by sloth

I'm not really impressed with the theories associated with quantum physics anymore.

Not really sure what that means. Sounds like you're confusing philosophy, which is largely speculative, with science, which is set in the stone of observation. You can't really say with a straight face you're not impressed with quantum physics when the very medium you're using to say that is the electronic devices for which quantum mechanics is indispensable.

It's an inherent fact of the universe (regardless of any 'theories' you might have about it) that there is an exact limit on the amount of knowledge you can have on small scales, and this can be used to create random devices that are, pretty much by definition really, perfectly random.

Originally Posted by Xei

Not really sure what that means. Sounds like you're confusing philosophy, which is largely speculative, with science, which is set in the stone of observation. You can't really say with a straight face you're not impressed with quantum physics when the very medium you're using to say that is the electronic devices for which quantum mechanics is indispensable.

It's an inherent fact of the universe (regardless of any 'theories' you might have about it) that there is an exact limit on the amount of knowledge you can have on small scales, and this can be used to create random devices that are, pretty much by definition really, perfectly random.

Where is this stone?
Where is the stone that said that meat turned into maggots?

Yes, I realize that I am using a machine that was created by theoretic concepts based upon quantum theory. ...Doesn't change a thing.
So you say that someone else says that they believe that it is an inherent fact of the universe that there is an exact limit blah blah blah and they are BY DEFINITION random. I wouldn't want to insult Webster's power, would I...

It is a good thing that there are these people to tell us what to believe. You're not religious, by chance, are you?

Originally Posted by Dianeva

8 5 5 1 9 2 9 9 0 9 2 2 6 4 6 1 2 5 2 9 1 9 1 1 1 7 0 1 6 2 9 8 8

How did you think through that? Or did you try not to think? I feel like just looking at it...it seems very nonrandom.

8...88

19...29...92...29...19...29 - see what I mean?
And there's a lot of repeats for this short of a sequence.

seven 9's
only one 7
and no 3's

Maybe I don't fully understand the concept of random number generator?

Here's my attempt after pressing a bunch of keys in about a second:
3 1 9 0 8 4 5 7 6 3 8 2 4 7 9 5 6 4 0 1 3 4 5 2 6 3 1 5 4

Originally Posted by Aquanina

How did you think through that? Or did you try not to think? I feel like just looking at it...it seems very nonrandom.

8...88

19...29...92...29...19...29 - see what I mean?
And there's a lot of repeats for this short of a sequence.

seven 9's
only one 7
and no 3's

Maybe I don't fully understand the concept of random number generator?

Here's my attempt after pressing a bunch of keys in about a second:
3 1 9 0 8 4 5 7 6 3 8 2 4 7 9 5 6 4 0 1 3 4 5 2 6 3 1 5 4

It is strange that there were no 3's, I didn't notice that. I said that was my attempt, I wasn't claiming that I think I can simulate a random generator perfectly. I was wondering whether a person can or not.

And the problem with yours is that there are no repeats, while I think if each number were completely random it would be really unlikely to go through 25 or so numbers without a repeat.

Originally Posted by Dianeva

I said that was my attempt, I wasn't claiming that I think I can simulate a random generator perfectly. I was wondering whether a person can or not.

Oh I know that, I wasn't trying to imply that you had made any such claim. I just wanted to point out what I saw when I looked at your sequence. People probably make pretty bad random number generators. It's a really interesting question though.

Originally Posted by Dianeva

And the problem with yours is that there are no repeats, while I think if each number were completely random it would be really unlikely to go through 25 or so numbers without a repeat.

Yeah...it's funny...when I stuck in one or two repeats, the number started to look very nonrandom to me so I took them out. But you're right.

7 and i think 3 are special numbers for some reason and like you can them more or something


46193569125457429756148759350913425803276487013658 70327086428756134208754281563427561703428580342568 034265807326

Okay, nice observations. I don't think I would have noticed the 29, 19 thing myself. I guess I tended to hit one side of the keyboard and then the other.

There is actually a lot of behavioral data on this question. This research is of great interest to me but I'll try not to ramble for too long. The answer is definitively no, humans are generally very poor at simulating truly random sequences. We are also poor at identifying truly random sequences. The most common error that people make when attempting to generate a random sequence is that they underestimate the number of "runs" that would be expected by chance, that is, the likelihood that the same number will be repeated multiple times in a row. In human-generated sequences, runs tend to be both too infrequent and too short. Notice, for example, that out of the three human-generated sequences which have been presented in this thread so far, two of them contain not a single run of even two numbers in a row. This is actually incredibly unlikely, statistically speaking, but that's what looks "random" to a typical human. The third contains a few runs, even a run of 3 numbers, but as others have pointed out, there are some biases in the overall frequency of particular numbers. Usually these biases are less prominent than the bias against long runs.

Psychologists in the 1970s gave these statistical intuitions about the stability of samples the tongue-in-cheek name "belief in the law of small numbers": the erroneous belief that the law of large numbers also applies to small numbers. The law of large numbers holds that as the size of a sample increases (in this case, the total sequence of numbers grows larger), the results (the frequency of each number 0-9) approach their expected values (1/10 for each number). But people have the intuition that this law, which concerns the global sequence, ought to also apply locally to any subset of the global sequence: so for example, if we have a sequence of 100 numbers, people expect that the frequency of numbers in any given 10-number subset drawn from that sequence should also be pretty close to their expected values. But that's wrong; the law of large numbers only applies to large numbers! It applies to the global sequence but it is decreasingly true for smaller subsets of the sequence. This type of reasoning causes us to err when we generate random sequences because after we list a particular number, we get the feeling that we shouldn't list it again right after, because that would make the local sequence conform less well to the expected values--in other words, it would violate the "law of small numbers."

Yo dub, analyze me bro:

Code:

6 8 4 3 6 8 7 4 3 6 8 7 3 1 5 8 7 5 8 7 7 8 4 8 2 1 2 1 4 6 5 1 0 6 1 5 6 4 2 1 6 3 7 6 7 9 4 1 2 9 7 
6 2 4 1 9 7 3 2 4 1 9 7 2 9 8 2 9 7 4 2 3 9 7 2 9 2 9 2 4 1 9 6 7 2 9 1 6 2 7 9 1 2 3 7 9 4 1 3 9 1 6 4 3 2 
9 4 3 2 1 9 4 1 3 9 4 1 2 9 4 1 2 9 4 2 1 4 6 1 3 9 0 2 1 9 0 2 9 1 0 5 1 9 0 1 6 0 2 9 0 4 1 9 2 1 7 1 5 1
 4 0 3 5 1 4 0 6 1 6 1 8 0 1 6 8 0 1 6 8 0 4 6 8 0 1 6 0 1 6 0 8 1 3 0 1 6 0 4 1 0 6 8 4 1 6 0 8 7 0 4 1 6 6 
5 0 4 4 1 0 5 0 1 4 0 0 1 5 4 0 6 1 0 8 5 4 0 1 0 5 0 4 0 5 1 0 5 0 4 1 0 3 0 5 4 1 0 3 0 4 1 6 0 5 4 3 0 6 
5 4 1 0 5 1 0 3 2 4 0 6 5 0 1 3 0 2 1 3 0 5 1 6 0 5 1 6 0 5 1 4 0 6 5 1 0 3 0 2 1 0 6 5 1 0 6 3 0 2 1 0 6 5 
1 6 0 1 4 6 0 1 6 0 8 4 0 5 7 0 6 5 7 0 6 7 0  9 7 7 0 9 7 0 4 6 1 0 6 0 1 6 8 0 4 9 0 1 0 6 0 8 4 0 6 5 4 1 
0 6 5 5 1 0 6 0 1 0 6 6 8 0 4 0 6 8 7 4 0 6 8 4 1 9 0 1 0 5 1 6 8 0 4 9 0 8 4 7 0 6 8 1 0 6 8 1 0 6 8 7 1 0 
6 8 0 1 6 8 0 1 6 0 1 3 0 2 1 0 3 1 0 6 8 0 4 6 2 1 0 6 0 1 6 0 1 6 1 0 6 4 1 0 9 8 1 0 4 9 1 9 8 0 1 9 0 7 
1 4 9 0 8 4 7 0 9 8 7 4 0 9 8 4 1 0 6 0 1 6 8 7 4 0 9 1 8 7 4 9 1 0 8

edit: I just mashed this out on the number pad.

According to my analysis of your random number sequence, you have an irrational fear of public restroom facilities stemming from a repressed sexual memory from your early childhood.

How does that make you feel?

One of these sequences has been randomly generated with a tiny Java program I just made, the other is my attempt at randomness. Can you guess which is which? I'm going to make one first, and then I'll run the program and stick with the sequence generated on the first run. Then I'm going to make another mini-program to tell me which order I should list them in here. Try to guess which one I made and which one was generated:

1) 9 6 5 1 0 3 2 4 7 7 8 0 1 5 9 8 2 2 5 3 0 9 2 6 4 8 2 7 5 7

2) 9 4 6 8 4 3 4 1 3 8 9 1 9 2 7 8 5 5 7 9 5 0 4 0 3 0 3 1 7 8

It's a trick question- you can't program. You're a woman!

Code:

> seq1
 [1] 9 6 5 1 0 3 2 4 7 7 8 0 1 5 9 8 2 2 5 3 0 9 2 6 4 8 2 7 5 7
> runs.test(seq1)

	Runs Test - Two sided

data:  seq1 
Standardized Runs Statistic = -1.0951, p-value = 0.2735

> seq2
 [1] 9 4 6 8 4 3 4 1 3 8 9 1 9 2 7 8 5 5 7 9 5 0 4 0 3 0 3 1 7 8
> runs.test(seq2)

	Runs Test - Two sided

data:  seq2 
Standardized Runs Statistic = -1.8581, p-value = 0.06316

Sequence 2 is less likely to occur by chance.

"trying to be random" has its problems.

apparently magicians, gamblers, etc. have been using this flaw of predictability in humans as well. 'Pick a random number between 0 and 10', and the typical answer would be three or seven. The same with 'Pick a random number between 0 and 100', and the same thing would occur: 37 apparently is the most common answer.
I've heard of similar stories where some teacher asked his class to pull out their calculators and push the 'random' button 100 times or so to produce either a 'heads' (1) or a tails (0) result, and write down what they got. When he got the results, it is said that he actually could see who used the random number generator and who just went on ahead and *tried* to be random by skipping the time-consuming work of actually pushing the button and thinking of what they thought to be random strings of 0s and 1s. The people who tried to be random, for example, didn't have strings such as 1111111111111100 or something, which (apparently) is completely plausible when it is generated randomly.
We all know this.. We have all done multiple choice exams where at some point we got a huge line of answers that were completely the same. "what? BBBBBBBBBBB? That can't be right.."

Also: look up the Pilot episode of the Numb3rs series. It deals with this same subject. Plus, the series is pretty cool
http://en.wikipedia.org/wiki/Pilot_(Numb3rs)

Here's some other stuffs for you folks to look at, it's pretty interesting .

Another forum-based experiment:
xkcd • View topic - Human random numbers

Don't know how accurate this is, haven't found any papers to back it up, though I have some other info as well further down...
Jyte - Human-generated random numbers are usually odd. - Cast your vote

Remember the numbers 7 and 37 apparently being the most common human-generated random number? Here's something akin to that:
Is 17 the "most random" number? : Cognitive Daily

I've found some research surveys and papers as well:
http://wexler.free.fr/library/files/...literature.pdf
http://www.springerlink.com/content/...2/fulltext.pdf
http://www.ieice.org/proceedings/NOL...paper/5060.pdf
http://www.pharma.uzh.ch/research/ch...p_Res_2006.pdf

Also: some fun exercises to see whether you can behave randomly
Exercise Introduction


FWIW,

TimB

I do not believe humans are capable as functioning as decent RNG systems because of the various biases in play. If I remember correctly, many actually try and avoid 'patterns' such as double digits. For instance, people would tend to say HTHHTHTT is more random than HHTTHHTT.

The methodologies people use when trying to write a 'random' sequences are also flawed. If I'm typing on a keyboard, I might favour using a certain digit to type, restricting several characters to a certain range of digits due to limits on my range of motion. My input methodology might involve me moving around the keyboard from area to area which will not produce a random result. I might deliberately try to make sure something is 'random' enough and in doing so make it non-random.

Even if I asked someone to give me a digit every few days unprompted, hopefully after they forgot which one they last gave me, I suspect the person in question would be likely to pick certain numbers preferentially.

1) 9 6 5 1 0 3 2 4 7 7 8 0 1 5 9 8 2 2 5 3 0 9 2 6 4 8 2 7 5 7

2) 9 4 6 8 4 3 4 1 3 8 9 1 9 2 7 8 5 5 7 9 5 0 4 0 3 0 3 1 7 8

Edit: It would have helped if I actually hadn't misread and misinterpreted DuB's post (I blame being sleep deprived at the time) and so got the significance tests the wrong way round. But my mistake might still have biased my conclusion so I'll leave the rest of this as it was originally posted.

With a completely 100% scientific analysis* which conclusively proves my case as scientific fact** the second one appears to have a greater degree of freedom, thus more random. It looks as if the first one was made by someone varying the area of the keyboard in which they typed periodically, resulting in chains of characters from the same region of numbers.

*In other words, I just looked at them.
**Warning: may not be true

Though I could also be biased because I read DuB's post first, and indeed it might also be the case that he suspected this, and switched the results around so he could come along later and claim how this influenced my opinion. Or MAYBE he suspected I would realise this might be the case and therefore correctly claimed #2 is more random, expecting a skeptical person like me would figure this and change to selecting #1. Or perhaps he figured out I would deduce this and therefore incorrectly claimed #2 is more random, relying on people using higher order reasoning to come to the wrong conclusion. Or...

[several hours later]

...it could be the case where he realised that I would come to the conclusion that he realised that I suspected that he figured that I would think that he noticed that I decided that he thought that I would decide he was lying where he said #2 was less likely by chance, and so #1 is actually less likely.

*slowly descends into a quivering incoherent mess*

Or I'm using confirmation bias to point out what I already believe to be true and both are equally random and this is an experiment you two set up to investigate confirmation bias.

OR several forum members here decided to work on a project and discussed about how best to psychologically manipulate me whilst me being convinced of their non participance

*runs screaming from the room*

Originally Posted by Photolysis

Though I could also be biased because I read DuB's post first, and indeed it might also be the case that he suspected this, and switched the results around so he could come along later and claim how this influenced my opinion. Or MAYBE he suspected I would realise this might be the case and therefore correctly claimed #2 is more random, expecting a skeptical person like me would figure this and change to selecting #1. Or perhaps he figured out I would deduce this and therefore incorrectly claimed #2 is more random, relying on people using higher order reasoning to come to the wrong conclusion. Or...

Clearly you have a dizzying intellect...

Sometimes I read these Extended Discussion threads thinking I'll learn something of value, then I stop to think everyone's wasting their time pondering over shit that doesn't even matter, and then I head back to The Lounge feeling as if I'm not quite ready to understand the intricacies of the minds of ED-lurkers.

Oh woe is me, my simpleminded self forever stuck on the shores of the brainy sea...

Well, back to my corner.

Originally Posted by GavinGill

Sometimes I read these Extended Discussion threads thinking I'll learn something of value, then I stop to think everyone's wasting their time pondering over shit that doesn't even matter, and then I head back to The Lounge feeling as if I'm not quite ready to understand the intricacies of the minds of ED-lurkers.

Oh woe is me, my simpleminded self forever stuck on the shores of the brainy sea...

Well, back to my corner.

And yet a simple google search would yield this ;p

Applications of randomness - Wikipedia, the free encyclopedia
Randomness - Wikipedia, the free encyclopedia

Statistics, sciency stuff like testing medicines, human psychology, quantum mechanics, simulation, SECURITY ;p
Hmm..

Then again.. Does anything have value apart from what we humans make of it?

etc etc

A few people have mentioned all of the biases. But would it be possible for a person to learn about their biases, until a point where they really can simulate a random number generator because they're aware of their biases and know how to avoid them?

In an infinite sequence, a random number generator is going to make every combination eventually, even "1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1". So any sequence a person makes up will be a valid randomly generated sequence, one that a generator could have generated. So is the answer just yes, that people can simulate a random number generator, even considering bias? Every sequence has the same probability of being generated. The generator is just as likely to generate the sequence "1 2 3 4 5 6 7" as it is "3 9 1 1 7 3 6".

In a randomly generated sequence, as the sequence's length approaches infinity, there should be almost the same amount of each number. '7' might have been generated 8092 times, '6' generated 8011 times, etc. But if a human with a bias were to try it, any preference for a particular number would show up and the numbers would be more like 7: 9055, 6: 5090. Unless the person has a further bias to address this problem, and is more likely to purposely try, after long enough, to cover numbers he didn't before more often.

I'm more confused about this topic than I was before.

In the two sequences I posted above, it was (1) that I made up and (2) that was generated, so DuB's program's result was wrong, and Photolysis was right.

Originally Posted by GavinGill

Sometimes I read these Extended Discussion threads thinking I'll learn something of value, then I stop to think everyone's wasting their time pondering over shit that doesn't even matter, and then I head back to The Lounge feeling as if I'm not quite ready to understand the intricacies of the minds of ED-lurkers.

Oh woe is me, my simpleminded self forever stuck on the shores of the brainy sea...

Well, back to my corner.

Some of the topics probably are pointless, apart from being entertainment to those who find them interesting. They shouldn't be limited to people who have a lot of knowledge about the subjects (I definitely don't. I've never even heard this topic brought up before). If you don't find pointless philosophical topics interesting, there's no shame in not taking part in them.

Originally Posted by Dianeva

So is the answer just yes, that people can simulate a random number generator, even considering bias?

Well sure, if we're just talking about the trivial sense in which it is possible for a human and a randomizer to produce the same sequence, then of course. But it seems like the more interesting question is: can humans naturally and consistently simulate a randomizer? And the answer to that is definitely No.

Originally Posted by Dianeva

In the two sequences I posted above, it was (1) that I made up and (2) that was generated, so DuB's program's result was wrong, and Photolysis was right.

Damn... probability theory is bullshit!

Originally Posted by Dianeva

Some of the topics probably are pointless, apart from being entertainment to those who find them interesting. They shouldn't be limited to people who have a lot of knowledge about the subjects (I definitely don't. I've never even heard this topic brought up before). If you don't find pointless philosophical topics interesting, there's no shame in not taking part in them.

I think there are enormous practical implications to the question of whether or not people accurately perceive randomness. When we "see" apparent patterns when in fact there is nothing but the vagaries of chance--we see something when in fact there's nothing--we develop inaccurate mental models of the world around us which systematically lead us astray, with potentially serious consequences. For example, placing bets according to erroneous beliefs such as gambler's fallacy or streak shooting in basketball will cause you to systematically lose money. And when we fail to see the real patterns that are in front of us because they appear random--we see nothing when in fact there's something--we systematically neglect information that we could be using to make more optimal decisions. In short, whether or not humans deal well with randomness is a Big Deal.

Damn... probability theory is bullshit!

Ah well, if it makes you feel better, I misread your post this morning like an idiot when I mistakenly believed you reckoned #2 was more random. Apparently I got distracted by pretty numbers.

Ooh, shiny.

A few people have mentioned all of the biases. But would it be possible for a person to learn about their biases, until a point where they really can simulate a random number generator because they're aware of their biases and know how to avoid them?

I think one of TimB's links mentioned something about how people can act more random with training.