Distributiontype

Exact odds for a spec, worked out rather than simulated: every total the dice can make and how likely it is.

Plain dice are counted. Ten d1000 is a million million million million million outcomes, far past what a number holds exactly, so the counts are kept as BigInt, whole and exact however large, and only turned into a probability at the end. exactCounts hands those counts out. Rerolled and exploding dice, and several dice kept from a pool, are not equally likely outcomes to count, so they are worked out as probabilities instead and are right to the last few digits a number holds. Nothing here is sampled or rounded off, and nothing is left out: exploding dice stop at MAX_EXPLOSIONS, in the roll and in the odds alike, so the odds are those of the dice as they are really thrown.

import type { Distribution } from "@johnmorrisdotca/korokoro";

type Distribution = {
    min: number;
    max: number;
    probabilities: number[];
};

import type { Distribution } from "@johnmorrisdotca/korokoro";
import { DEFAULT_SPEC, distributionOf } from "@johnmorrisdotca/korokoro";
const value: Distribution = distributionOf(DEFAULT_SPEC);
console.log([value.min, value.max]);

[ 2, 12 ]