Question on Randomness

HigsBoson

Was having a conversation with a friend last night about random number generators. He says they always have something that the number, or process for getting the number, is based on, for example, in online poker they use the tinest fraction of a second in which the mouse was clicked as a starting point. He claims this cannot be truely random because the program is basing it on something, but if you use tracking software you can see that the more hands you play the closer you get to an even distribution. My question is, in the human brain do we do the same? If I am asked to pick a number between 1 and 1000 and pick 725, did I do something similar to the program? I think it would make an interesting study, get a lot of people to randomly pick numbers and see if you get an even distribution, then see if there is a difference between individual and group results. Does anyone know of such a study?

cavedave

Humans cannot pick random numbers. 17 is the most random number to humans.

You can make a true random number generator using physical devices. Radiocativity is the canonical example of a ramdom process (As far as we know)

Any one who considers arithmetical methods of producing random digits is, of course, in a state of sin

Von Neumann.

There is a fun seven state mind reading machine here based on Shannon's work for anyone who thinks they are random.

djpbarry

HigsBoson wrote: »

If I am asked to pick a number between 1 and 1000 and pick 725, did I do something similar to the program? I think it would make an interesting study, get a lot of people to randomly pick numbers and see if you get an even distribution...

I doubt it - I imagine very few people would choose 1 or 1,000, for example. You'd possibly get a skewed Normal distribution centred around about the 500 mark.

HigsBoson

Cheers CaveDave, will read those links tonight, looks very interesting.

djpbarry wrote: »

I doubt it - I imagine very few people would choose 1 or 1,000, for example. You'd possibly get a skewed Normal distribution centred around about the 500 mark.

Is that only if each person chooses one number? Would it be the same if they had to spend hours on end coming up with numbers?

djpbarry

HigsBoson wrote: »

Is that only if each person chooses one number? Would it be the same if they had to spend hours on end coming up with numbers?

Ah, you mean have a series of individuals acting as random number generators? I imagine in that case there would be a large degree of repetition for each individual - I'd be very surprised if a uniform distribution resulted, but that's purely a hunch.

Capt'n Midnight

cavedave wrote: »

Humans cannot pick random numbers. 17 is the most random number to humans.

ask your friends to write down a random number between 1 and 4 odds on they will choose 3 , try it for yourself


have a look at http://random.org if you want some info or a few random numbers












http://www.boallen.com/random-numbers.html

Random.org



old PHP random

Waking-Dreams

I'm currently reading 'The Drunkard's Walk: How Randomness Rules Our Lives' by Leonard Mlodinow, which is pretty good and relates to the OP.

HigsBoson

Waking-Dreams wrote: »

I'm currently reading 'The Drunkard's Walk: How Randomness Rules Our Lives' by Leonard Mlodinow, which is pretty good and relates to the OP.

Had this in my hand in Easons the other day but bought another book, will pick it up at the weekend. Cheers.

amacachi

cavedave wrote: »

There is a fun seven state mind reading machine here based on Shannon's work for anyone who thinks they are random.

Got it below 50% a few times, bloody hard though. :pac:

EDIT: If I leave it ages between clicks so I don't know what my last click was I can fairly easily get it below 50%. I'd imagine it would be the same if a few people were taking turns and didn't know the last click since it goes by patterns.

cavedave

It is only a few state machines. If you think about it you can beat it easily enough. It is just odd that something you can write in less than a dozen lines can do a good job of predicting you.