Palworld Capture Rate Explained: Full Formula, Shake Checks, and Bonus Rules
The capture reticle shows a percentage, but many players notice their actual catch rate does not match that number. The reason is simple: the reticle uses its own display scale, while a real throw runs a complete formula plus three shake checks. This guide takes the whole mechanism apart, verified line by line against version 1.0 internal capture logic, so you can treat it as the authoritative reference.
The companion capture rate calculator already implements every rule in this article. Pick a Pal, drag the sliders, and you get exact probabilities for 10 sphere types across 6 scenarios.
The Complete Formula
The base capture probability P of a single throw is computed in four steps.
Step 1: level difference L
- Sphere level: a fixed strength value per sphere type, see the table below.
- Capture power level: raised through the Statue of Power. Version 1.0 has 15 levels, each converting to +0.5, up to +7.5 at max.
- Passive bonus: capture levels granted by the partner skills of certain Pals in your party. Wispaw's "Death-Cheating Feline" grants up to +5 while the Back Bonus is active, Muffly adds a similar bonus against frozen targets, and Souffline does the same against entangled targets.
- Target level: the wild Pal's level. Rare (shiny) Pals are counted here as
max(level − 5, 1). - Weak point hit: when the sphere lands directly on the target's weak point (the head for most Pals), the sphere level for that throw counts as +2. The occasional "+X%" badge next to the reticle displays exactly the probability gap those 2 levels create.
| Sphere | Level value | Sphere | Level value |
|---|---|---|---|
| Pal Sphere | 7 | Legendary Sphere | 38 |
| Mega Sphere | 14 | Ultimate Sphere | 44 |
| Giga Sphere | 20 | Exotic Sphere | 50 |
| Hyper Sphere | 27 | Sol Sphere | 58 |
| Ultra Sphere | 33 | Ancient Sphere | 64 |
Mental math for L is easy: sphere level plus the capture power conversion (+1 per 2 levels), minus the target level. The sign alone tells you whether the encounter is worth a throw. Three examples:
| Scenario | L | Starting point |
|---|---|---|
| Lv20 target, Giga Sphere (20) + capture power Lv10 (+5) | +5 | Positive level difference, a good start |
| Lv45 target, Ultimate Sphere (44) + capture power Lv5 (+2.5) | +1.5 | Barely positive; workable with HP reduction and ailments |
| Lv45 target, Giga Sphere (20) + capture power Lv5 (+2.5) | −22.5 | Deep negative, basically a wasted throw |
The last two rows are the scenarios used in the full example below: the same Lv45 Jetragon with L moving from −22.5 to +1.5 swings the probability by nearly three hundred times. L itself is only an intermediate quantity. How it turns into a probability is step two.
Step 2: the level factor nl
L ≤ −50: nl = 0
otherwise: nl = (L + 50) / 99
nl folds L into a factor between 0 and 1. Plotted, it is a ramp capped at both ends, with no staircase jumps:
The middle of the ramp is a straight line: 1 point of L is worth roughly 0.01, and both ends cap at ±50. A few numbers to build intuition:
| L | nl | Situation |
|---|---|---|
| Below −50 | 0 | The level gap is hopeless; the base probability is plain zero and no status effect can save it |
| −25 | 0.25 | Sphere and bonuses clearly under the target; only HP reduction and ailments keep it alive |
| 0 | 0.51 | Level value exactly cancels the target level, starting from one half |
| +25 | 0.76 | Combined level value outweighs the target by 25, already high probability territory |
| Above +50 | 1 | Capped; a larger level difference brings no extra benefit |
Cross reference with the full example below: a Giga Sphere against a Lv45 Jetragon gives L = −22.5 and nl only 0.278, and after the ninth power compression the probability is nearly zero. Swap in an Ultimate Sphere and L turns positive at 1.5, nl climbs to 0.520, and the probability immediately enters usable range. Every bit the level difference moves toward the positive is sharply amplified by the ninth power that follows. That is the mathematical root of the "pick the right sphere first" tip.
Step 3: the HP factor and the shape function
A ≥ 0.5: P₀ = 1 − (2 − 2A)⁹ / 2
A < 0.5: P₀ = 256 × A⁹
A is a single strength value that merges the level difference (through nl) and remaining HP, and the shape function maps it to the base probability P₀. The formula alone is not intuitive, so just look at the curve:
Read the curve in three segments:
- Dead zone (A < 0.35 or so): the ninth power flattens small values to almost nothing. At A = 0.3, P₀ is only about 0.5%; at A = 0.263 (the Giga Sphere versus Lv45 Jetragon scenario) only 0.15% remains. Within this segment the relative gains from HP reduction and ailments are decent, but the base is so low that a throw is still wasted in practice.
- Steep slope (A ≈ 0.35 to 0.55): the steepest stretch of the curve. Each 0.05 of A multiplies the probability by two to three (6.7% at A = 0.40, 19.4% at 0.45, 50% at 0.50). The practical border between "worth a throw" and "not worth it" sits right here, and the Ultimate Sphere from the example above (A = 0.492) lands mid slope at 43.6%.
- Saturation zone (A > 0.6): the curve quickly presses against the 100% ceiling: 93% at A = 0.6 and 99.5% at 0.7. Past this point, stacking more level difference yields nearly nothing, so save the higher spheres for harder targets.
About the split itself: the two formulas are exactly equal at A = 0.5 (both equal 0.5), and even the slopes match (9 on both sides), so this is not two rules stitched together but one smooth curve written in two pieces. The low branch lets weak throws fall cleanly toward zero, the high branch lets strong throws converge gently to 1, and the junction is completely invisible.
The HP term 1.3^(−HP%) means the factor is only 0.77 at full HP and rises to 0.97 at one tenth remaining. Its job is to slide A along the horizontal axis: the lower the HP, the further right A moves, deeper into the steep slope and the high probability zone.
Step 4: the outer multipliers
- Status multiplier:
1 + 0.3 × ailment count + 0.35 × sleep. Poison, Burn, Soak, Freeze, Shock, Mud, Ivy, and Darkness each add +0.3, sleep adds +0.35, stacking linearly. Note that in the game most ailments are mutually exclusive: a Pal can carry only one elemental ailment at a time, and only Poison can coexist with another ailment; sleep runs on an independent flag and does not compete with ailments. The theoretical stacking cap is therefore Poison plus one ailment plus sleep, ×1.95 in total, but in practice you can almost never assemble it: no active skill, partner skill, or item can inflict sleep. It only appears on targets that fell asleep naturally (usually at night), and any damage or breakout from the sphere wakes them instantly. The two realistic routes, pick one: sneak a direct throw at a full HP sleeping target at night (×1.35), or Poison plus one more ailment in open combat (×1.6). The restraint from bear traps is the strongest single status bonus (+0.5, i.e. ×1.5), but almost nobody uses bear traps in the game, so ignore it in real play. - Species multiplier: every Pal carries a fixed capture difficulty factor, read straight from game data. Common early Pals such as Lamball and Cattiva sit at 1.5, while legendary Pals such as Jetragon and Xenolord sit at 1.0.
- World multiplier: the PalCaptureRate (capture rate multiplier) setting on custom servers, 2.0 on easy, 1.0 on normal, 0.8 on hard, adjustable from 0.5 to 2.0; see the server settings tool.
The final P is clamped between 0 and 1.
Three Shake Checks: Why the Displayed Value Is the Actual Value
After landing, the sphere shakes three times. Each shake is an independent check, with pass rates of:
The three exponents add up to exactly 1, so the total probability of passing all three is P itself:
The widely shared claim that "actual probability = displayed value^(1/3)" comes from a reasonable but wrong assumption: that all three checks share the same probability. In reality the three check lines are all different, and their product equals the displayed value exactly.
One more trap that is easy to step into: the in game reticle reading is not P itself. The reticle takes the single pass rate of the second shake and magnifies it once more (P^(1/3) raised to the 1.25 power, P^(5/12) in total), so the reticle always reads higher than the real probability. When the reticle shows 70%, the real P is only about 44%. For exact numbers, trust the formulas or the capture calculator.
A concrete number: at P = 43.6%, the three shakes are 69.1%, 75.8%, and 83.1%, multiplying back to 43.6%, while the reticle shows about 70.8%.
Not every throw has to survive all three checks. When any of these conditions holds, the first shake is an automatic pass:
With the first shake skipped, the actual probability of a single throw rises from P to P^(5/9) (only the last two checks multiply). The exponent 5/9 is smaller than 1, so whenever P < 1 the transformation always raises the probability: 43.6% becomes about 63.0%, and 10% becomes about 27.8%. The HP condition deserves special attention: in real play you almost always lower a target below half HP before throwing, so most actual throws carry a probability higher than P. The first shake skip is the most easily overlooked piece of the whole formula. The sneak route has one extra gate: the target itself must be in a sneak-attackable state, which in practice means unaware or naturally asleep targets; once a fight has started, a sphere landing squarely on the back no longer counts as a sneak attack. Past that gate it applies to every sphere: all 12 sphere types carry the full sneak parameter, from the Pal Sphere to the Ancient Sphere alike.
Four Widely Shared Claims
| Claim | Reality |
|---|---|
| Actual probability is the cube root of the displayed value | False. The three shake exponents sum to 1, so the total probability equals the displayed value exactly |
| Sneak attacks from behind give a capture bonus | True. A sphere that connects from behind on a sneak-attackable target (unaware or asleep) skips the first shake, raising the single throw probability from P to P^(5/9). With Wispaw in the party, "Death-Cheating Feline" also converts the Back Bonus into up to +5 capture levels added to L |
| Multiple ailment bonuses multiply together | They stack linearly: 1 + 0.3 × ailment count + 0.35 × sleep. Elemental ailments are mutually exclusive, so the practical maximum is Poison plus one elemental ailment, ×1.6 not ×1.69 |
| Bosses carry a ×0.7 penalty in the code | Mostly true. Boss entries carry a lower species multiplier, roughly ×0.7 of the regular version, so estimate boss catches with that discount |
A Rare Pal's +5 and −5
Rare (shiny) Pals spawn 5 levels above their kind, yet the formula counts them 5 levels lower. The two corrections cancel exactly: within the same encounter, a shiny is no harder to catch than a regular one. Compare a shiny with a regular Pal of the same level and the shiny is even slightly easier.
A Full Worked Example
Target: a Lv45 Jetragon (species multiplier 1.0), lowered to 20% HP, capture power Lv5 (+2.5), normal difficulty.
With a Giga Sphere (20): L = 20 + 2.5 − 45 = −22.5, nl = 0.278, A = 1.3^(−0.2) × 0.278 = 0.263, landing in the low branch, P = 256 × 0.263⁹ ≈ 0.15%. Essentially a wasted throw.
Swap to an Ultimate Sphere (44): L = 1.5, nl = 0.520, A = 0.492, P ≈ 43.6%. Same Pal, same HP bar, and one sphere tier multiplies the probability by nearly three hundred.
Now add Poison and Shock: 43.6% × 1.6 ≈ 69.8%. That is the real magnitude of every multiplier in the formula.
This final throw: 69.8%^(5/9) ≈ 81.9%. The target is already below 50% HP and the capture rate is at least 50%, so both threshold routes are met: the first shake is skipped automatically, no matter the angle. Using the sphere, the ailments, and HP control all in one fight lifts the probability yet another notch.
Verify with the Calculator
The capture rate calculator ships with the species multipliers of all 299 Pals and the level values of all 10 spheres. It supports Pal level, remaining HP, capture power, the world multiplier, and a rare toggle, and each sphere card lists the exact single throw probability for regular and boss Pals across three scenarios: standard, ailment, and sleep. After reading this article, drag a few sliders in the calculator and the sense of scale will click immediately.