nerd sniped myself even further and made a rust crate for droptables
no idea why, just some recreational programming before bed. its fast!
gn nostr 😴
I was randomly thinking about loot/drop table implementations in games and naturally started thinking of the fastest way to implement large drop tables.
gpt5 pointed me to this fascinating algorithm that is barely mentioned online, but it can turn sample any discrete probability distribution in O(1)
https://y7k4.github.io/2021/03/23/alias-method.html
its a bit brain bending and it took a good 20 minutes of staring to understand how it works, but its pretty cool.
#nerdsnipe
View quoted note →
if it takes 500 boss kills on average for something to drop, you can just specify that drop rate and then when you sample the table it will roll it for you in O(1).
You can have values that don't add up to 1, it all gets normalized.
You can even have it roll multiple tables. one would roll rarity, then you could have it roll all the legendary items if it hits the legendary loot table. you can keep rolling this way.
very common thing in RPGs, so was fun making this library.
