How breeding actually works in Palworld

Breeding in Palworld doesn't average your two Pals into a third — it uses a hidden "breeding power" number attached to every species, and the egg that comes out is whichever Pal sits closest to the midpoint of your two parents' numbers. Put a Chikipi next to a Frostallion and you don't get a strong hybrid; you get whatever middling Pal happens to occupy that spot on the ladder. That's the one thing that has to click before anything else about breeding makes sense.

What you need before the first egg

Breeding needs four things in place at once: a Breeding Farm, unlocked at level 19 in the Ancient Technology tree; a Cake sitting in the farm's storage box; one male and one female Pal of breedable species placed inside; and an Egg Incubator nearby to actually hatch what comes out. Skip the cake and nothing happens — no egg, no error message, the two Pals just sit there. The full ingredient list for the items database if you want to check what a Cooking Pot needs before you commit a play session to farming it.

Cake isn't one item, either. Palworld 1.0 ships five variants, and which one you use changes what the egg does, not just whether you get one:

Special Cake is the expensive one, and it's the one worth saving for once you're breeding for passives rather than just for a species you're missing.

The power level: why the "obvious" pairing doesn't work

Every breedable Pal has a breeding-power value baked into it, and it's this number — not the parents' levels, stats, or how strong they are in a fight — that decides the child's species. The two parents' values get combined and rounded, and the game looks for the closest species match on that combined value. That means the result isn't "somewhere between mom and dad" in any visual or elemental sense: two fire Pals can produce something with no fire typing at all, if that's what lands nearest the number.

This is also why a breeding combinations database is the only sane way to plan a specific target — it already lists the 181 known parent pairs and what each one produces, so you're not hand-averaging hidden numbers to reverse-engineer a result.

Exclusive combos: the pairs that skip the formula entirely

A separate set of Pals — mostly elemental subspecies, like the "Lux," "Noct," "Cryst," "Ignis" and "Aqua" variants — don't come from the power-level formula at all. Each one has one fixed, specific parent pair, and breeding that exact pair always produces that specific Pal, regardless of what the power-level math would otherwise say. Fuack + Flambelle always gives Fuack Ignis; Pengullet + Jolthog always gives Jolthog Cryst. If a target Pal isn't reachable by adjusting power level, check whether it's one of these first — you'll save yourself a lot of wasted eggs. As of version 1.0, none of these variants are breeding-exclusive: every one of them can also be caught in the wild, breeding is just the reliable way to get one on demand. The full list of pairs is in the same breeding database.

Passive skills: what actually carries over

Passive skills are the other half of why breeding matters — a Pal can hold up to four, and combat, work, and mount Pals all want different ones. When two Pals breed, the offspring's passives are drawn from the pool of passives its parents are actually carrying, plus a chance of a passive that neither parent had. Both sources agree there's no guarantee: even a passive both parents share can simply fail to appear in the child, and it's entirely possible to hatch an egg with none of the passives you were breeding for. Special Cake and the tools that improve talent development narrow the odds in your favor, but they don't remove the randomness — which is why serious passive breeding usually means hatching several eggs from the same pair, not one.

The practical approach is chain breeding: pick the target passive, find a pair of Pals that each carry it (or carry passives that complement it), breed repeatedly, and keep the eggs that land the combination you want, working your way up through better carriers each generation rather than expecting one lucky hatch. Checking the Pals database for which species can even carry your target passive is the step people skip and then wonder why their pairing never turns anything up — not every Pal can hold every passive.

What this means for planning a breeding project

Decide the target first — species or passive, not both at once on your first attempt — then work backward. For a specific species, the breeding database tells you the exact pair, or whether it's one of the fixed exclusive combos instead of a power-level result. For a specific passive, pick a carrier pair, budget for several hatches rather than one, and only bring out Special Cake once you're past the "which two Pals" stage and into the "how many attempts" stage — it's the ingredient list, in items, that decides whether that's a five-minute detour or an afternoon at the Ranch and Plantation.