<Greys>
egg, can you think of any combination of dice that when summed together create a probability distribution which is an asymmetric bell curve?
<Iskierka>
rather than combination, method of combination? Multiplying them should approach a maxwell-boltzmann distribution
<Iskierka>
or at least the same kind of asymmetry, probably not the same actual shape
<Iskierka>
other than that any dice with an asymmetric distribution should do that (1, 1, 1, 2, 2, 3 on sides) but I don't think you can by adding up regular dice
<Greys>
hmmm, multiplication should be workable, with a sufficiently high ratio, say d20 * d4
<Greys>
this distribution actually seems to be too irregular
<Iskierka>
I believe the drop highest/lowest method also gives irregular, though I don't know by how much
<Greys>
1d3+(1d3*1d3) looks pretty good, if I could reverse it
<Greys>
this would be used in a FATE like system that normally uses 2d6; I don't love that because 7 is significantly more probable in a 2d6 roll, and you need to roll at least an 8 to get what you want
<Greys>
what does this syntax on output 2 mean?
<Iskierka>
access items 2 and 3 from the collection of 3 dice, which are ordered from highest to lowest, so is effectively "drop 1 highest"
<Iskierka>
use 1,2 to drop lowest and invert the curve
<Greys>
inverted is best, and this is an easy roll to perform
<Greys>
great, so then the players are more likely to succeed
<Greys>
roughly 1/3rd chance of failure, almost 30% chance of mediocre succeess, more than 1/3rd chance of full success, and 20ish% chance of critical success
<Iskierka>
This is why D&D uses the drop lowest for stats, and effectively also does for advantage, though it words it differently
<Greys>
any idea on how to smooth that down a bit?
<Iskierka>
Not hugely without irregular dice, unless you just want to use a fixed +- number to reduce crit or something
<Greys>
I just want to make the min and max likelihoods slightly closer together, maybe get the min to 2% and bring the max down to more like 14%, with the same proportional shape
<Greys>
but this works
<Greys>
d8 is a bit nicer
<Iskierka>
http://anydice.com/program/c17d might be a bit aggressive and makes 1 possible but is another option with any even-sided dice
<Greys>
{3}@(d8) seems wrong
<Greys>
maybe I'm not good at statistics
<Iskierka>
I'm not sure what result that should return as you're asking for the third die of a collection of exactly 1 d8
<Greys>
{3}@(3d8)
<Iskierka>
That looks right to me; it's selecting the lowest of 3 d8s
<Greys>
right, this is the probability of the other two being equal or higher than N
<Iskierka>
of rolling N and the other two being equal or higher
<Iskierka>
which does make it somewhat different
<Greys>
so there's a .2% chance in rolling 3d8 of getting 3 8s
<Iskierka>
!wa (1/8)^3 as decimal
<Qboid>
Iskierka: N[(1/8)^3] = 0.001953125
<Iskierka>
yep
<Greys>
eh, maybe I don't need to think about this