If a card has a 0.0043% chance to appear in one slot, what are the odds of seeing it at least once in a five-card pack? And how many packs would give you a 50% chance? Here is a worked example using GRIEF from Extreme Demon Gacha.
Start with the chance of missing
Let p be the chance of drawing the target card in one slot, expressed as a decimal. Assume each slot is independent and uses the same odds. The chance that all five slots miss is (1 - p)^5. Subtract it from 1 to get the chance of at least one hit:
Chance per pack = 1 - (1 - p)^5
The same idea works across any number of packs. For n five-card packs, there are 5 * n slots, so the chance is 1 - (1 - p)^(5 * n).
A worked example: GRIEF
In Demon Index’s November 18, 2026 data snapshot, GRIEF’s per-slot rate was about 0.00429682%, or 0.0000429682. That gives about 0.02148% per five-card pack. After 100 packs, the chance of seeing at least one copy is about 2.13%.
| Target chance | Packs needed |
|---|---|
| 50% | 3,227 |
| 90% | 10,718 |
| 95% | 13,944 |
At 3,227 packs, roughly half of independent runs would still end without a copy. If you only want a holo and the holo roll is an independent 10%, multiply the per-slot rate by 0.10 before calculating the pack odds.
Calculate it in JavaScript
const perSlot = 0.00004296819632637584; const slotsPerPack = 5; const logMissPerPack = slotsPerPack * Math.log1p(-perSlot); const chanceAfter = packs => -Math.expm1(packs * logMissPerPack); const packsFor = target => Math.ceil(Math.log1p(-target) / logMissPerPack); console.log((100 * chanceAfter(100)).toFixed(2) + "%"); // 2.13% console.log([0.5, 0.9, 0.95].map(packsFor)); // [3227, 10718, 13944] Math.log1p and Math.expm1 help preserve precision when the per-slot rate is very small. The model assumes independent, fixed-rate slots. Guaranteed slots or pity mechanics need their own model.
Try a different card or pack count in the fan-made Demon Index odds calculator. The methodology and data sources explain how the tool uses the game’s published rates.