Randomness and the human mind/body

[curiousnoncat]

I'm posting this here because it's neither fully statistics or biology, and I have no technical knowledge of statistics.

It's well known that humans have trouble behaving truly randomly. If a group of people is asked to choose a random number between 1 and 10, each person is very likely to choose seven. The group may actually behave randomly, but with a normal distribution.

I'm not interested in group behavior though, I was wondering if I can generate a seemingly random sequence using only my mind or a part of my body, no dice, no coin-flips, etc.

I closed my eyes and typed a sequence of 990 1's and 0's, but there's only about 43% 1's, and when you scale it to this size a sort of chaotic pattern becomes apparent.

If you compare it to the random bits from Random.org on the right, you will see the difference.

My bits probably wouldn't meet a statistical definition of randomness, because the "run length" of 1's is likely all screwed up.


I'm forced to rule out conscious effort as a good generator of random data, so I was wondering if there might be a random element to the timing of blinking, or some other biological process. Please help me by brainstorming which elements of human behavior are most "random", or some method of imitating the unpredictability of dice rolls, ideally with my hands. Just out of curiosity.


http://www.newgrounds.com/dump/draw/4c8e6eefd8138159e414a16bcde54dd5

 

[fargoth]

Well, in unix-like operating systems, the "\dev\random" random generator "device" uses an entropy pool that, among other things, takes into account user's mouse movements and keystrokes (e.g. time between keystrokes).
This data might still not be quite random enough (i.e. with high enough entropy), so a nifty algorithm is used to determine which bytes to take from the pool (or pools, in case it's something like Fortuna).

If Roger Penrose is correct, then the firing of your synapses is a random process (that seems odd to me, since we're pretty predictable. I, for example, never surprise myself by my actions).
If, on the other hand most other people are correct, then we're just chaotic systems (probably deterministic) - in which case, true randomness is beyond us :)
But the algorithms I was talking about are pretty good at imitating true randomness from our behaviour.

 

[curiousnoncat]

But the algorithms I was talking about are pretty good at imitating true randomness from our behaviour.

I believe I have heard about that. However, that requires a computer program designed specifically for detecting such things. I'm mostly interested in things that can be done by monitoring yourself or performing some action analogous to a dice roll, using only the mind and body.

I'm sort of imagining being locked up nude in a jail cell, with nothing around you, and being told you can be set free if you can say aloud a sequence that fits the requirements for randomness.

 

[zoobyshoe]

Isn't a deliberate effort to be random automatically non-random?

 

[fargoth]

Well, since you only "run" once, you could pick a large number, and apply some pseudo random algorithm on that number.
You don't really need a computer for that... just concentrate real hard :)
If you'd use "Mersenne Twister" it would take you more than 10⁶⁰⁰⁰ random seeming numbers to repeat the pattern.
pretty good for a naked dude in a prison cell - you would probably need to live an unnaturally long life to say that many numbers :)

 

[curiousnoncat]

Isn't a deliberate effort to be random automatically non-random?

There are, as fargoth said, random elements to human behavior, I just want to know some clever ways to observe them.

Here's a bad example: If you could lift your arm and let it fall back down onto a surface, and use its position to create a random number.

However, that probably wouldn't be very random, because the tension in your muscles, and the starting position would be pretty predictable, and you would probably just repeat a pattern of actions.

 

[fargoth]

How about variations in cardiac rhythm? http://link.springer.com/content/pdf/10.1007%2F978-3-0348-8119-7_21.pdf#page-1
or http://proceedings.aip.org/resource/2/apcpcs/285/1/697_1?isAuthorized=no
and finally a free access article: http://people.clarkson.edu/~ebollt/Papers/Heart_v04.pdf

 

[fargoth]

There are, as fargoth said, random elements to human behavior, I just want to know some clever ways to observe them.

Here's a bad example: If you could lift your arm and let it fall back down onto a surface, and use its position to create a random number.

However, that probably wouldn't be very random, because the tension in your muscles, and the starting position would be pretty predictable, and you would probably just repeat a pattern of actions.

That's putting words in my mouth :)
I've said I don't believe we are truly random (in the sense of e.g. a radioactive decay).
I think we are just chaotic systems (and hence stochastic in the sense of deterministic statistical mechanics).
In the sense you're talking about "randomness" it amounts to the same observable behaviour, since we can never re-run the same system twice and watch what happens (unless time travel will become feasible at some point - in which case true randomness would decree that you would most probably act differently in each rerun, while my "randomness" would predict you'd do exactly the same thing again).

 

[curiousnoncat]

How about variations in cardiac rhythm? http://link.springer.com/content/pdf/10.1007%2F978-3-0348-8119-7_21.pdf#page-1
or http://proceedings.aip.org/resource/2/apcpcs/285/1/697_1?isAuthorized=no
and finally a free access article: http://people.clarkson.edu/~ebollt/Papers/Heart_v04.pdf

That seems to have good potential, if there is a timepiece in the room, or the individual can count a consistent 'unit' of mental time, not necessarily seconds. I guess once you found the average rate then you would just measure the small variations from the average.

 

[DevilsAvocado]

What could be more random than your raw brain activity?

For less than a hundred bucks you get access to the stochastic stuff by the http://www.neurosky.com/Products/MindWave.aspx headset that among other things measures EEG:

https://www.youtube.com/watch?v=1tr4CjtGtvg


And there are lots of free/commercial applications & hack like this one:

https://www.youtube.com/watch?v=Rqgd5TykxAc


And if this isn’t enough, you can even measure REM Sleep activity!

https://www.youtube.com/watch?v=um6t85LLUH8


If you do programming, there’s also a free .NET library http://brianpeek.com/post/thinkgearnet-1-1.

 

[curiousnoncat]

What could be more random than your raw brain activity?

For less than a hundred bucks you get access to the stochastic stuff by the ...

Interesting.

I was just brainstorming before I saw your post, and what I have thought of is a much simpler electronic way to get random digits.

I think if a person weighs themselves with a high accuracy scale with many digits after the decimal, and then drinks some water or eats some food, the weight of that material is somewhat indeterminate, so the digit or digits on the smallest order of magnitude would be sufficiently random.