Pick a monster and the kind of kill — champion, unique, minion, quest, super-unique, Terror Zone, herald — and read what one kill drops, item by item, with Magic Find and players; or name an item and find the monsters that drop it best.


The chance one kill drops each (any quality for a base); a unique or set item is its base × the quality check × its share among the uniques / set items that fit the item level.derived
| # | Kill | |||
|---|---|---|---|---|
| 1 | Quest | 88 | 1 in 1,965 | |
| 2 | Quest | 87 | 1 in 2,055 | |
| 3 | Quest | 94 | 1 in 2,136 | |
| 4 | Quest | 99 | 1 in 2,166 | |
| 5 | Normal | 87 | 1 in 3,189 | |
| 6 | Normal | 94 | 1 in 3,310 | |
| 7 | Normal | 99 | 1 in 3,359 | |
| 8 | Normal | 88 | 1 in 5,278 | |
| 9 | Super-unique | 95 | 1 in 27,766 | |
| 10 | Super-unique | 86 | 1 in 28,389 | |
| 11 | Minion | 86 | 1 in 28,389 | |
| 12 | Super-unique | 88 | 1 in 41,611 | |
| 13 | Super-unique | 88 | 1 in 41,611 | |
| 14 | Super-unique | 88 | 1 in 41,611 | |
| 15 | Super-unique | 88 | 1 in 41,611 | |
| 16 | Super-unique | 88 | 1 in 41,611 | |
| 17 | Super-unique | 88 | 1 in 41,611 | |
| 18 | Super-unique | 88 | 1 in 41,611 | |
| 19 | Super-unique | 85 | 1 in 41,789 | |
| 20 | Super-unique | 86 | 1 in 41,789 | |
| 21 | Super-unique | 86 | 1 in 41,789 | |
| 22 | Super-unique | 86 | 1 in 41,789 | |
| 23 | Super-unique | 86 | 1 in 42,773 | |
| 24 | Super-unique | 88 | 1 in 44,589 | |
| 25 | Super-unique | 86 | 1 in 44,781 |
Every monster and kind of kill on Hell at its own level; heralds left out. A click opens that kill.
Every run is an independent try at the same odds: a miss never makes the next run likelier. The mean is where the average farmer lands; half need more than the median, one in ten more than the 90 % figure. Your run time is yours to measure — the game does not record it.derivedderived
One kill rolls 7 picks from Baal (H). Shares are within their parent; an opened TC's meta line gives its odds per root pick.
c = (ratio − (item level − base level) / divisor) × 128; then × 100 / (100 + effective Magic Find); at least the row's minimum; then less c × the TC's quality factor / 1024. The check succeeds with 128 / c, in the order unique, set, rare, magic. A Shako from Hell Mephisto at 0 % Magic Find is about 1 in 14.9 per check.
Magic Find counts in full for magic items, and for every quality up to 10 %; above 10 % it has diminishing returns for the others: MF × 250 / (MF + 250) for unique, × 500 / (MF + 500) for set, × 600 / (MF + 600) for rare.
The expansion rows apply. The Uber row is for exceptional and elite bases, the class-specific row for class items; the monster's rank does not choose the row.
With n players N = (n + 1) / 2, rounded down; the no-drop weight becomes total / ((no-drop + total) / no-drop)^N − 1, floored.
On Nightmare and Hell a monster's treasure class moves down the rows that follow it in the table while they are in its group and their level is at most the monster's level. Bosses never upgrade (monstats boss, NoRatio), nor does anything on Normal; a Terror Zone's Desecrated classes always do.
weap, armo, mele, bow and abow groups (3 to 96) hold every spawnable item of that kind whose level, rounded up to a multiple of 3, is the group's number — no quest items, no throwing potions — weighted by its item type's Rarity (at least 1), not the item's own.
A negative pick count runs the entries in order, each as many times as its weight, with no no-drop roll — and stops when the weights run out, even with picks left.
One kill drops at most six items; later picks are lost.
A treasure class's Unique, Set, Rare and Magic factors pass down to the classes it calls; where two meet, the higher one holds. The lab carries them to every item, so a Mephisto drop reached through his 983 factors keeps them.
An item's level is the level of the monster that drops it: champions 2 above their type, uniques, their minions and super-uniques 3 (a super-unique at its home area's level + 3 on Nightmare and Hell). In a Terror Zone the monster is first raised to your level + 2 within the zone's bounds; a herald adds its tier's treasure-class boost. The treasure class upgrades at that level.
Runs an hour = 3600 ÷ seconds a run. A run's chance = 1 − (1 − per kill)^kills of the target in the run. The expected runs to the first drop is 1 ÷ that chance — a mean, not a promise: half of all farmers need more than the median run count, and 90 % and 99 % take the runs shown. The chance in an hour or in ten is 1 − (1 − per run)^runs.
Each run is independent and the odds never improve after a miss: the chance of at least one drop in n runs is 1 − (1 − p)^n, and n for a given confidence c is ⌈ln(1 − c) ÷ ln(1 − p)⌉.
Per pick, each item's chance is the product of the shares down its path; a unique or set item multiplies its base by the quality check (with the factors in effect) and its share among the uniques or set items that fit the item level. Per kill, the picks run in the game's order — a sequence's entries one after the other, a nested class's picks in a row — and an item counts when it shows among the first six drops.
A unique roll picks among the base's spawnable uniques whose level is at most the item level, weighted by rarity; with none it becomes a rare (a set roll becomes magic).
Sources: treasureclassex.txt, itemratio.txt, itemtypes.txt, weapons / armor / misc.txt, uniqueitems.txt, setitems.txt, monstats.txt, superuniques.txt, levels.txt and desecratedzones.json of build 3.3.93847 (data/tools, pipeline/build-tools-lab.mjs).
Per kill = the chance it shows among the first 6 drops of 7 picks. legacy rule
Fury Visor · qlvl 66 · exceptional / elite (Uber row) · item level 99
| Magic Find | Unique |
|---|---|
| 0 | 10.9% |
| 50 | 15.4% |
| 100 | 18.6% |
| 150 | 21.0% |
| 200 | 23.0% |
| 250 | 24.5% |
| 300 | 25.7% |
| 350 | 26.7% |
| 400 | 27.6% |
| 450 | 28.3% |
| 500 | 29.0% |
| 550 | 29.5% |
| 600 | 30.0% |
| 650 | 30.5% |
| 700 | 30.9% |
| 750 | 31.3% |
| 800 | 31.6% |
| 850 | 31.9% |
| 900 | 32.2% |
| 950 | 32.3% |
| 1000 | 32.7% |
| Magic Find | Set |
|---|---|
| 0 | 20.4% |
| 50 | 28.1% |
| 100 | 34.0% |
| 150 | 38.9% |
| 200 | 42.7% |
| 250 | 42.9% |
| 300 | 42.3% |
| 350 | 41.7% |
| 400 | 41.2% |
| 450 | 40.8% |
| 500 | 40.4% |
| 550 | 40.1% |
| 600 | 39.8% |
| 650 | 39.6% |
| 700 | 39.3% |
| 750 | 39.1% |
| 800 | 38.9% |
| 850 | 38.7% |
| 900 | 38.6% |
| 950 | 38.5% |
| 1000 | 38.3% |
| Magic Find | Rare |
|---|---|
| 0 | 24.8% |
| 50 | 29.8% |
| 100 | 31.6% |
| 150 | 31.9% |
| 200 | 31.0% |
| 250 | 32.3% |
| 300 | 31.8% |
| 350 | 31.4% |
| 400 | 31.0% |
| 450 | 30.7% |
| 500 | 30.4% |
| 550 | 30.2% |
| 600 | 29.9% |
| 650 | 29.7% |
| 700 | 29.6% |
| 750 | 29.4% |
| 800 | 29.3% |
| 850 | 29.1% |
| 900 | 29.0% |
| 950 | 29.0% |
| 1000 | 28.8% |
| Magic Find | Magic |
|---|---|
| 0 | 43.9% |
| 50 | 26.8% |
| 100 | 15.8% |
| 150 | 8.22% |
| 200 | 3.39% |
| 250 | 1 in 396 |
| 300 | 1 in 403 |
| 350 | 1 in 408 |
| 400 | 1 in 413 |
| 450 | 1 in 417 |
| 500 | 1 in 421 |
| 550 | 1 in 424 |
| 600 | 1 in 428 |
| 650 | 1 in 430 |
| 700 | 1 in 433 |
| 750 | 1 in 436 |
| 800 | 1 in 437 |
| 850 | 1 in 440 |
| 900 | 1 in 441 |
| 950 | 1 in 442 |
| 1000 | 1 in 444 |
The chance a dropped base ends each quality: the checks run unique → set → rare → magic until one succeeds; Magic Find counts in full for magic and with diminishing returns for the rest.derived