RandomChances - Easy to use (ToS compliant!) random chances library

This is a really simple library I made that handles “random chance” events/item selections for you. Its easy to use, ToS complaint, and is unbiased (has inbuilt “security” to prevent seed guessing attacks/etc).

Example:

local Rand = require(script.RandomChances)

local Res = Rand.Run(
{
	--Each item has to add up to 100% in total. The amount you put in is the percentage you actually get such item.
	["epic item"] = 20,
	["even epicer item"] = 20,
	["even epicer-er item"] = 20,
	["super epic item"] = 30,
	["omega epic item"] = 5,
	["MOST epic item"] = 5,
})

print(Res) -- "epic item" (for example, will change every time you call it)

Some statistics of using this:

local Count = {}
local Events = 
{
	["epic item"] = 20,
	["even epicer item"] = 20,
	["even epicer-er item"] = 20,
	["super epic item"] = 30,
	["omega epic item"] = 5,
	["MOST epic item"] = 5,
}

for I, V in pairs(Events) do
	Count[I] = 0
end

for I = 0, 100000 do
	local Res = Rand.Run(Events)
	
	Count[Res] = Count[Res] + 1
end

for I, V in pairs(Count) do
	print(I, V)
end

--[[
Output:

epic item 19835
even epicer item 20659
even epicer-er item 19851
super epic item 29590
omega epic item 4992
MOST epic item 5074
]]

The library itself uses the Random API - the best option we currently have for random numbers. While it isn’t very secure, its all we got so I attempted to use it in the best way possible.

Each chance is set into a translation table that specifies the percentage bounds of each item/event. For example, 5 random items of 20% each would be split into the following entries:

Entry 1: 0-20
Entry 2: 20-40
Entry 3: 40-60
Entry 4: 60-80
Entry 5: 80-100

After that, it generates a random number from 1-100 and then searches through the translation table to see which entry it corresponds to.

The RNG is re-seeded every 1-75 iterations with the current time (the seed that Roblox itself uses) and the last random number generated to hopefully provide some more entropy.

Link to module. Enjoy!

22 Likes

Hmmm, I really want to test this. Can’t we just use rarity weights and calculate the percentage based on rarityWeight / totalWeight?

4 Likes