Palworld Breeding Calculator – What Do Two Pals Make?

target = floor( (power A + power B + 1) ÷ 2 ) → nearest breedable power

The famous one. Penking (520) + Bushi (640). Target = (520 + 640 + 1) ÷ 2 = 580. Anubis sits at 570 and Incineram at 590 — both exactly 10 away. Anubis comes first in the game files, so the egg is an Anubis, every single time.

A tie you can test. Anubis (570) + Cattiva (1460). Target = (570 + 1460 + 1) ÷ 2 = 1015. Felbat (1010) and Robinquill (1020) are both 5 away; Robinquill is earlier in the files, so the child is a Robinquill.

When a recipe overrules the maths. Vanwyrm (660) + Anubis (570) averages to 615, where the formula alone would give Incineram. But this pair is one of the fixed recipes, so the egg is a Faleris — the recipe is checked before the averaging ever happens.

Why you cannot breed your way to a legendary. The target is an average, so it can never be lower than your lower parent’s power. Two common parents can creep toward rare territory over generations, but a Jetragon (90) only ever comes from two Jetragon — or from the wild.

The default breeding — Penking at breeding power 520 with Bushi at 640 — gives a target of 580, from (520 + 640 + 1) ÷ 2 rounded down, and hatches Anubis at 570, the closest match. Every figure in this section was produced by changing the stated control on this page and reading the result back. The formula never changes; only the parents do.

The tie that is not a coin flip. The default target of 580 sits exactly between two Pals: Anubis at 570 and Incineram at 590, each 10 away. The game does not flip a coin — it settles the tie by game-file order, and Anubis, at file position 1, beats Incineram at file position 2. The next closest matches show how the ranking continues: Surfent at 560 is 20 away, then Elphidran at 540 is 40 away. Ties are common on a scale of 137 Pals spread over 1,490 points of range, because many targets fall exactly between two neighbours — and every tie on this page is resolved the same way, by the hidden file order the calculator applies for you. A breeder who expects a random outcome from a tied target will be surprised exactly half the time; a breeder who knows the file order will never be surprised at all.

Swapping the parents changes nothing. Swapping Penking and Bushi — Bushi first, Penking second — hatches Anubis again, at the same target of 580. Addition commutes, so parent order cannot matter to the formula, and the calculator honours that: there is no “mother” or “father” effect in the standard formula, no matter which slot a Pal sits in. That symmetry is worth stating because it contradicts the intuition many players bring from other breeding systems, where the mother’s species decides the child. Here the parents are two numbers and a sum, and the sum does not care which number was written first. The one place order does matter is the single gender-sensitive variant recipe — Wixen plus Katress, where which Pal is female decides whether the child is Wixen Noct or Katress Ignis — and that exception proves the rule, because it lives in the fixed-recipe list, not in the formula.

Same species, same child. Breeding Anubis with Anubis hatches Anubis: the target is (570 + 570 + 1) ÷ 2 rounded down, which is 570, and the closest Pal to 570 is Anubis itself at distance zero. That trivial case is the key to the whole legendary rule: Jetragon, Frostallion, Frostallion Noct, Paladius, Necromus, Shadowbeak and their kind are never produced by the standard formula — no mixed pair, however rare the parents, will hatch one — but two Jetragon together give a Jetragon, because the average of a Pal with itself is itself. The rarity scale runs from 10 (Blazamut, the rarest) to 1500 (Chikipi, the most common), and lower means rarer — so the child of any mixed pair is always an average, never an extreme, and the only way to hold an extreme is to breed it with itself.

Working backwards: 171 ways to Anubis. Switching to Child-to-Parents mode and choosing Anubis shows 171 parent pairs that produce it — Penking plus Bushi among them, alongside pairs as unrelated-looking as Rushoar plus Suzaku Aqua and Celaray plus Mammorest Cryst. That number is the formula’s real lesson: the child is a nearest-neighbour lookup on one number, so any pair whose average lands within Anubis’s catchment — closer to 570 than to any other Pal’s power — hatches Anubis, regardless of what the parents look like, what elements they carry, or how big they are. Two completely different pairs giving the same child is not a bug or a coincidence; it is the same sum landing where it lands. The reverse mode turns breeding from a guessing game into a shopping list: pick the child you want, and the page shows every pair that makes it, common parents first.

Each row is a breeding probed on this page. Powers are the game’s hidden breeding powers; lower means rarer.

The table’s shape is the whole game in miniature: averages, ties, symmetry, and one hard exception. Learn those four rows and every other breeding on the page is the same four ideas with different numbers.

The child is never rarer than the rarer parent. Because the target is the average of the two parents’ powers, and lower means rarer, the target always sits between the parents — never below the rarer one’s power. To move a bloodline toward a rare Pal, the page’s advice is to breed the rare parent with its own offspring again and again: each generation’s average creeps closer to the rare parent’s power, because the offspring is always nearer to the rare parent than the common parent was. That creep is slow by construction — an average moves halfway, not all the way — so a bloodline project is measured in generations, not in single breedings. A breeder who expects one lucky pairing to jump from common to legendary is expecting the average to do something averages never do.

Variants ignore the formula entirely. Each variant Pal has a single fixed recipe, checked by the calculator before any averaging happens: Mossanda plus Grizzbolt always makes Mossanda Lux, Jormuntide plus Blazehowl always makes Jormuntide Ignis, and so on through 31 fixed combinations. A variant’s breeding power in the table — Mossanda Lux at 390, Jormuntide Ignis at 315 — describes where it sits on the scale, not how it is made, because it is never made by the scale. The one gender-sensitive recipe is the exception inside the exception: Wixen plus Katress makes Wixen Noct when the Wixen is the female, and Katress Ignis when the Katress is the female. If a breeding you expected from the average gives you a variant instead, the recipe list ran first — and if a breeding you expected from a recipe gives you something ordinary, the recipe’s conditions, including gender where it matters, were not met.

What the data covers, and what it does not. The breeding powers, the tie-break order and the fixed recipes are the game-file values published in the Palworld Wiki’s breeding tables, captured in October 2026, across all 137 Pals in the calculator’s table. Game updates can renumber Pals or add new ones, and the page says openly that four documented variant recipes whose parents were added in later updates are not in its table yet. That omission is stated rather than hidden, and it is the right way to handle a living game’s data: a breeding calculator that silently omitted recipes would send breeders chasing pairs that cannot work, while one that states its gap lets them check the wiki for exactly those four cases. For everything else in the table, the figures are the game files’, not estimates — which is why the calculator can give an exact answer instead of a probability.

Elements, size and level do not enter into it. Nothing about elements, size or level enters the formula — it is pure arithmetic on one hidden stat, which is why a sheep and a cat can make a fox and why two unrelated pairs can make the same child. That purity is also what makes the reverse mode trustworthy: because the child depends only on the average, every pair that averages into a child’s catchment genuinely produces that child, and the list of 171 pairs for Anubis is complete for the table as captured, not a sample. The breeding farm’s surprises are therefore never mysteries, only sums landing where they land — and once a breeder thinks in powers rather than in species, the whole farm becomes legible. CalcFolk is a fan-made tool and is not affiliated with Pocketpair; the game itself, not this page, is the final word on any breeding.

The creep principle from the previous section sounds abstract until it is worked with real Pals, so here it is in three generations, all probed on this page. Generation one is the default: Penking at 520 with Bushi at 640 gives a target of 580 and hatches Anubis at 570. Generation two breeds the child back toward the rarer parent: Penking at 520 with Anubis at 570 gives a target of 545 and hatches Elphidran at 540 — closer to Penking’s 520 than Anubis was, by 30 points of power. The bloodline has moved toward the rare end, not by a leap but by an average: the offspring of a rare parent and a mid-range child is always nearer the rare parent than the child was, because the average of 520 and 570 must sit between them.

Generation three shows the same principle in the other direction: Anubis at 570 with Bushi at 640 gives a target of 605 and hatches Incineram at 590 — moving toward the common end, toward Bushi’s 640, rather than toward rarity. The direction of the creep is decided entirely by which parent is rarer: breed toward the rare parent and the line creeps rare; breed toward the common parent and it creeps common. There is no way to breed “outward” in both directions at once, and no pairing whose average escapes the interval between its parents. A bloodline project is therefore a sequence of choices about which parent to breed back to, made one generation at a time, with the calculator showing each generation’s child before the next pairing is chosen.

Choosing among many pairs is itself a skill the reverse mode teaches. When a target like Anubis offers 171 pairs, the list is ordered with common parents first, and the right pair is rarely the first one — it is the pair you already own, or can catch soonest, or whose other parent is itself useful for the next generation. A pair that produces the child but consumes your only rare parent is worse than a pair that produces the same child from two common Pals, because the rare parent is needed again for the next creep step. Read the pair list as an inventory problem, not as a ranking: the best pair is the one that leaves your breeding stock strongest for the generation after this one.

The practical way to run such a project is to work backwards from the target using the reverse mode, then forwards through the generations. First, choose the Pal you want and read its 171-style pair list — every pair that produces it, common parents first — and pick a pair you can actually assemble. Then breed that pair, take the child, and breed the child back toward whichever parent sits nearer your ultimate target, checking each generation’s child on the page before committing the next pairing. The reverse mode tells you what is possible; the forward mode tells you what each step costs in generations. Together they replace the breeding farm’s folklore — lucky pairs, favoured species, element theories — with arithmetic that can be checked before an egg is ever laid. And when a generation surprises you, check the tie first: a target exactly between two Pals is settled by file order, not by luck, and the page shows the near misses so the surprise can be diagnosed rather than merely suffered.