Let's see if /tg/ knows dice. You have a fair d6. You

>>52194442
What probability distribution?

>>52194442
I proceed to ignore the dice and copy down from memory the statistics on what will happen.
>inb4 the brain was never a listed tool

>>52194442
I'm going to assume that we can use super annoying, long solutions.

1. Determine number of digits in die you're emulating
2. Roll 1d6, rerolling if you roll a six, multiply by 2. Treat a 10 as a 0.
3. Roll 1d6. If even, add add 1 to your prior result.
4. You now have a 0-9, emulating a d10. Repeat 2-3 for each digit of your target die. If you roll over your target for any given digit, simply reroll until you're not over.

>>52194522
Oh, and if a digit does not exceed 5, you can instead use a single d6, treating a 6 as a 0, rerolling if you roll over.

d359, roll d6's for first and second digit, 2d6 for third.

>>52194471
The d6 has a perfectly even distribution. It's got an exact 1/6th chance of each of 1, 2, 3, 4, 5, and 6.

The goal is to use it and/or the other stuff to emulate another perfect die with an arbitrary number of sides.

>>52194515
Brain is of course allowed, but how do you use the statistics to generate a number.

>>52194522
Ding! That works almost flawlessly!
You do, if I'm reading your algorithm right, suffer from some errors when not going for a dice that's a flat power of 100 or single digit, because some results are dependent in likelyhood on other results. For example, the odds of getting a 6 or higher on the 10s place of the d359 depend on if you got a 3 for the 100s place.

>>52194442
Just off the top of my head, I can get all numbers with prime factors 2 ([1,3]=1, [4,6]=2), 3 ([1,2]=1, [3,4]=2, [5,6]=3), and 5 (6 = 1d6, recursive), which is enough for all the "normal" dice. I can't figure out how to get any larger primes into the system, since all tables of results you can generate will be 6^n in size.

Alternatively, >>52194522, but it doesn't look as pretty.

>>52194442
>Can you use the available materials to do so, and if so how?
Sure, just write 999 results on the paper and drop one of the dice from a few feet above it.

What's the second die for?

File: Binary.jpg (287KB, 1284x1950px) Image search: [Google]

287KB, 1284x1950px

easy

1 to 3 = tails

4 to 6 = heads

>>52194442
Roll 1d6 4 times, treating 6s as 0, and writing down the result in sequence. If I rolled 6,2,3,4, then I would write down the number 0234, or 234. Reroll any result greater than 4344 or lesser than 2.
Convert the result from base 6 to base 10.

Done.

>>52194604
How do I use statistics? Start rolling like hell.

Bit a math there and there for eliminating remainders, and eventually I got myself a number.

But that takes a long time. And that's why I hope this thread is purely theoretical.

>>52194618
The trick is in just rerolling until you get a valid result. I'm not aware of any method that wouldn't require rerolls to trim the result set down for emulating something like a d7, but I'd love it if somebody posted one.

>>52194621
How do you ensure a good distribution?
Second die is because I'm a faggot who took all of five seconds to grab a stock image that happened to have two dice in it.

>>52194657
Ding! Works!

>>52194662
Ding! Works! Fukken nerd with your different base systems. That's the first way I thought of too.

>>52194722
>How do you ensure a good distribution?
I don't need to, according to your post. You just asked for a method to generate a dX result.

If you really want to, you can use some of the paper to create a box around the rolling surface and shake the construction vigorously for each roll.

>>52194814
Except any common die in use has an equal chance to get any of its possible results (clearly ignoring weighted/novelty dice). If you're being asked to emulate 1d12, I'd assume you're trying to get 1/12th odds for each.

>>52194442
I think to myself, "Why the fuck do I need the level of granularity that 2d500 generates?" Then I think fuck making FATAL 2: Anal Boogaloo and just run GURPS or WEG Starwars.

>>52194862
You're tacking on additional criteria after the fact. Nowhere in the OP does it specify that the results have to be fair, just that the physical d6 itself is mechanically fair.

>>52194442
Convert X to base 6, roll a d6 for each digit in base 6 (as you do with two d10s for a d100).

You might have to reroll for the leftmost digit.

it's the optimal way regarding the number of d6 used.

>>52194442
Do it in binary. Roll the D6 for each digit; evens are 0, odds are 1. Re-roll if you roll a number too high.

For instance: You need to roll a D27. That's 11011 in binary. Roll the D6 5 times (you'll need to re-roll if you roll 1-1-1-1-1).

>>52194442
>write 0-9 on ten different sheets of paper
>shuffle them
>draw
>reshuffle
>draw again
>etc
>if I manage to get a 0001 I eat the dice and kill myself

>>52194722
>>52195556
I thought of base 6 too. While there are some cery creative results here, surely base 6 us just the best answer