Best way to generate random numbers at home?

[muttrox]

@Sapo

I am sorry I didn’t specify. No dice, no computers. I mean common daily objects in your grandma’s house.

My grandma has dice and coins (and a calculator and a computer). Where are you drawing the line, I don’t understand your line.

[clairobscur]

@Sapo

Is this really the best way to generate random numbers with stuff generally found at home and without resorting to making a mess, using higher math, etc?

I used your prof’s method in a number of instances (I used to play RPG a lot). Except that I used the penultimate digit instead of the last one, since the last one being even or odd depends on the page you check (left or right).

[ultrafilter]

@Walton_Firm

That’s a bad method. People are very bad at avoiding patterns. There are interesting experiments wherein people have been asked to type “as randomly as possible” on a keyboard – it turns out to be not very difficult to write a computer program which can predict, with surprising accuracy, which key they will hit before they hit it.

Do you have a source for that? If so, we’d appreciate it over here.

[Thudlow_Boink]

@clairobscur

the last one being even or odd depends on the page you check (left or right).

Yes, this is one major flaw in using the OP’s method exactly as described.

I would think that using the last (or next-to-last) digit of “randomly chosen” phone numbers out of a phone book would be a better method than using “randomly chosen” page numbers—assuming that all ten digits are equally likely to appear as the last digit of a phone number, which seems reasonable but I don’t know for sure.

@Sapo

For example, if you need a random number from 1 to 47, you open the first page, that is your last digit (units). Repeat for the next digit (tens). Here you can map 6-0 as 1-5 or discard them. If you got a 48, then start from the beginning.

I don’t know what you mean by “map 6-0 as 1-5.” For your first digit, you could map 5-9 as 0-4 or discard them. For your second digit, you can’t do any such “mapping.” Even if you needed a random number from 1 to 44, it wouldn’t be fair to do such “mapping,” because then the numbers 40-44 would each have twice as many chances to appear as any of the others.

[Indistinguishable]

@muttrox

Look at the second hand of a clock. Divide the 60 seconds into equal segments, and see what you get. That gets you 2,3,4,6,12,15,20, or 30 choices.

For generation of just a few random bits, this idea seems the best to me. The only problem is, if you want to generate more random digits, you really can’t do so immediately; human bias as to how long to wait before re-looking will creep in, at a level significant in terms of quantizing the result into seconds.

How to solve this problem? Perhaps get a stopwatch with the ability to spit out time at an interval so fine-grained that you couldn’t possibly expect human bias to matter so far as the least-significant digit goes.

[Sapo]

@Thudlow_Boink

Yes, this is one major flaw in using the OP’s method exactly as described.

Come to think of it, you are right. I may be misremembering. This was years ago. Second to last number might have been it.

I would think that using the last (or next-to-last) digit of “randomly chosen” phone numbers out of a phone book would be a better method than using “randomly chosen” page numbers—assuming that all ten digits are equally likely to appear as the last digit of a phone number, which seems reasonable but I don’t know for sure.

This is a big assumption that I am not sure I am comfortable making. It would be great if someone in the telco industry had an answer. This might merit its own thread.
The stopwatch method I have used. Hundredths of a second is all it takes (at least for my slow reaction times) to be close enough to random.