How the maths works

Your rally's damage depends on how strong each troop type is and how many troops of each you send. Stronger types deserve a bigger slice of the rally. The maths below scores each type so the calculator can split your rally proportionally, then estimates the damage your hosted rally deals.

The dev's official name for the event is Bear Hunt; "Bear Trap" is the colloquial label. The building is the Pitfall, the boss is the Raging Bear.

The community-consensus formula, attributed to "Daryl from the Kingshot Discord" and reproduced by every serious guide site:

army_t = sqrt(N_t × 5000)

attack_per_troop_t = base_atk_t × (1 + Atk%_t) × base_leth / 100 × (1 + Leth%_t)

damage_per_round_t = army_t × attack_per_troop_t / (DEF × HP) × SkillMod_t

t

troop type: infantry / cavalry / archer.

N_t

number of troops of type t in your rally.

base_atk_t

datamined base attack at your troop tier (see §08).

base_leth

10 – fixed for every troop and every tier.

DEF, HP

the bear's stats: 10 defence, 83.33 health, giving an effective defence of 8.33.

SkillMod_t

the rally's hero-skill multiplier for that type (see §03).

Battles run for 10 rounds. Total damage is the sum across troop types, summed over rounds, ceilinged. Troops don't die against the Bear, so with no hero skills SkillMod is 1 and every round deals the same.

SkillMod is the part of the formula hero skills drive. The community decomposition, tested in-game:

SkillMod = (DamageUp × OppDefenseDown) / (OppDamageDown × DefenseUp)

The Bear has no buffs of its own, so the defender-side terms (DefenseUp, OppDamageDown) are 1 and SkillMod collapses to essentially DamageUp – the product of all attack and lethality skills in the rally.

DamageUp = product over effect_op ids of (1 + sum of same-id skill values / 100)

Every buff carries an effect_op id. Buffs with the sameid add together; buffs with different ids multiply. That single rule decides whether stacking heroes is worth it:

4× Chenko (all id 101): 1 + 100/100 = 2.00

2× Chenko + 2× Amane (id 101 + 102): 1.5 × 1.5 = 2.25

Mixing a Lethality hero with an Attack hero beats four of the same by about 12.5%, because the two ids multiply instead of add. Archers also carry a permanent ×1.10 from Ranged Strike, since the Bear counts as an infantry-type target.

For the ratio optimiser, each troop type gets a score:

S = (1 + Atk%) × (1 + Leth%) × base

Atk% and Leth% are your in-game per-type buffs (research, gear, charms, hero passives combined). base is the per-troop attack ratio across types – an exact 1 : 3 : 4 (see §06).

Worked example: Infantry with 200% Atk, 150% Leth, base 1 gives S = 3 × 2.5 × 1 = 7.5.

The calculator weights each type by S² when splitting the rally. Doubling a type's per-troop strength is "worth" four times as many troops, because troop count enters the damage formula under a square root.

From §02 a type's damage scales with sqrt(N). Maximising total damage:

maximise   D = Σ Sᵢ × √Nᵢ      subject to   Σ Nᵢ = M, Nᵢ ≥ 0

The Lagrangian gives Sᵢ / (2√Nᵢ) = λ, so √Nᵢ ∝ Sᵢ, hence Nᵢ ∝ Sᵢ². That is the closed-form optimum and the calculator applies it directly.

Per-troop base attack is datamined from State of Survival, which shares a battle engine with Kingshot (both Century Games / FunPlus). Defence and Lethality are fixed at 10 for every troop and every tier; only Attack and Health scale. Base attack at in-tier substep 0:

The Infantry : Cavalry : Archer ratio is exactly 1 : 3 : 4 at every tier – it holds at T1, T6 and T11 alike. So the recommended split never depends on your tier; only the absolute damage estimate does (see §08).

The recommended split is a broad plateau, not a knife-edge. This curve walks the archer share from 0 to 100%, holding Infantry and Cavalry at their optimal sub-ratio; damage is shown as a percentage of the best possible split.

Open the calculator and enter your buffs first.

Anything from 48% to 74% archers lands within 1% of the best possible damage. Get the split roughly right and stop sweating the decimals.

The Troop tier control (T1–T11, T11 = Truegold) sets which row of the §06 table the damage estimate uses. Because the 1 : 3 : 4 ratio is identical at every tier, changing tier never changes the recommended Inf/Cav/Arc split – it only scales the damage number to your real troops.

The Substep control (0–5) is the in-tier TrueGold refinement stage of your troop-training building – the Barracks, Stable or Range past building level 30. Each step adds roughly +4–5% Attack, so substep 5 is about +26% over substep 0. Like tier, it scales the damage estimate only and never shifts the recommended split. Read it straight off the building's TrueGold stage.

The Archer TrueGold skill toggle models the archer combat skill unlocked at troop refinement TG5 – an extra +25% archer damage against the infantry-type Bear, multiplied on top of the +10% Ranged Strike. Turn it On only if your archers are TG5 or higher. Single-source community estimate.

A rally's damage depends on all 13 hero skills in it: the leader's 9 (three heroes × three skills) plus up to 4 joiner first-skills. You control your own heroes; you do not control who joins your rally. So the calculator gives two numbers:

Floor – damage from your three rally-leader heroes' first skills only. The honest worst case: you host, nobody useful joins.

Ceiling – floor plus the four strongest first skills from your joiner pool, standing in for four strong alliance joiners. The realistic best case.

Valora's Hunter Instinct (§12) is applied on top of both. The estimate covers only the rally you host. Damage from joiner marches you send to other players' rallies depends on a host you don't control and is not modelled.

Rosa contributes two distinct effects to a rally she joins. Many calculators conflate them; this one keeps them separate.

Golden Rhythm (Expedition slot 3): "Increasing Archers' total Attack by 6 / 12 / 18 / 24 / 30%" across Lv 1–5. Archer-only, additive on Atk%.

Perihelion (Aeolian widget Expedition skill): "Rally Squads' Lethality 5 / 7.5 / 10 / 12.5 / 15%" across Lv 1–5. Applies to every troop type in the rally. Independent of Golden Rhythm; both can be active if the player owns the Aeolian widget.

The alliance's Pitfall building is upgradeable; each level adds +5% Attack to all troop types in the rally. At Lv 5 that's a flat +25%. Full upgrade costs 1,900 Hunting Arrows.

Because the buff is a flat percentage, it shifts the ratio toward whichever types have the lower baseline Atk% – adding 25% to a type at 50% Atk is proportionally bigger than to a type at 200%.

Two pet skills stack onto the same Atk / Leth slots as Pitfall (§11) and feed straight into the per-type score S in §04. Same effect-op grouping: identical-id values add within their bucket, distinct ids multiply across buckets.

Wild Charge (Giant Rhino) adds Squad Attack % across every troop type. Ladder Lv1–Lv10: 2.5 / 3 / 3.5 / 4 / 5 / 6 / 7 / 8 / 9 / 10 %.

Deadly Bite (Alpha Black Panther) adds Troop Lethality % across every troop type. Same ladder.

Because both buffs are uniform across Inf / Cav / Arc, neither shifts the optimal ratio – they just scale the whole damage number. Lethality is harder to stack from gear and heroes than Attack, so Deadly Bite typically moves the needle further at endgame even though the % is identical.

Valora is a Bear-Hunt specialist Master. Her Hunter Instincttalent adds Bear Hunt Damage Points on a non-linear ladder:

It is a top-line multiplier applied after the per-type formula, on your own damage score only. Because it scales every type equally, it never shifts the recommended ratio – it only moves the final number.

Valora also has two capacity talents – Dance of the Hunt(+30,000 per level to Raging Bear Rally Squad Capacity) and Savage Advantage (+3,000 per level to Bear Hunt Squad Capacity). The calculator does not take these as inputs: they are permanent Master talents, so the capacity the game shows you already includes them. Adding them again would double-count. Just enter the cap as the game displays it. Fearless Roar is different – it is a temporary activated pet buff, so the calculator does add that.

Cassia is the fourth Master (released 25 May 2026), tagged "battle-focused". Unlike Valora, she has no Bear-Hunt-specific damage multiplier – her kit is conventional Squad Atk / Def / Lethality / Health and Rally Capacity buffs. The calc takes no Cassia input: every relevant number she contributes is already in the in-game stat sheet and capacity you type in.

Her ledger and how the calc treats each entry:

Don't double-count. All five combat-relevant entries are permanent passives that the game folds into your visible Atk%, Lethality%, and march cap. The calc reads those visible numbers directly. Adding a separate Cassia panel on top would count her twice. Same precedent as Valora's Dance of the Hunt and Savage Advantage talents (§12, §16) – permanent Master buffs stay out of the calc inputs and live inside the cap the game shows.

Class neutrality. None of Cassia's buffs are troop-type- specific. They don't shift the optimal Inf/Cav/Arc ratio – they scale every type's score identically, so the §06 / §07 optimiser output is unchanged.

The army term sqrt(remaining × 5000) shrinks fast below 5,000 of any troop type. Five community sources agree on the same absolute floor: keep at least 5,000 infantry across your whole rally (not per march). The "10% per march" rule you'll see in chat is shorthand – the load-bearing number is 5,000 absolute.

Composition shifts across hero generations:

The 1/10/89 split is ratio-only; the absolute floor still applies. If 1% of your march cap is fewer than 5,000 troops, bump infantry up until you clear the floor.

Per the official Helpshift Combat FAQ:

rally skills = leader's full 9 + 4 first-skills from joiners

The rally leader contributes all 9 of their three heroes' skills. Joiners contribute only the first skill of their first hero, and only the top 4 joiner skills are picked across the rally. Selection rule: highest skill level wins, ties broken by join order. A high-level junk skill displaces a low-level damage skill – so joining with no hero beats joining with a junk hero.

Heroes stack by their effect_op id (§03): same id adds, different ids multiply. So 2× Chenko + 2× Amane – a Lethality hero paired with an Attack hero – beats 4× Chenko by ~12.5%.

Stacking happens within a single rally (1 leader + up to 4 joiners), not across the different marches you send. Pairing only pays off when you host and can dictate which heroes the joiners bring.

Up to September 2025 Bear Hunt also paid an alliance-wide reward tier topping out at 1.2 billion total alliance damage. That tier was cancelled in the 4 September 2025 update and consolidated into personal damage rewards; the 4 March 2026 update added further personal achievement rewards based on your historical total damage. The optimisation target is your own rally damage, not a shared pool.

Guides that still cite a "1.2 billion alliance cap" predate this change. The cooldown ends at UTC 00:00 on the second natural day after participation – day-aligned, not a fixed 48-hour timer.

Separate from damage: how many troops each march holds, and how the calculator fills them. Each march's working capacity is:

main_cap     = in-game cap + Fearless Roar + Antler Impact

joiner_cap   = in-game cap + Fearless Roar

Fearless Roar (Mighty Bison pet skill) adds +1,500 per level, max +15,000 at Lv10. Buffs Squad Capacity – every march in the rally, including joiners. It is a temporary, activated buff – 2-hour duration, 20-hour cooldown – so the capacity the game shows you while it is inactive does not include it. The calculator adds it, and only counts if you activate the pet skill in-game before marching.

Antler Impact (Great Moose pet skill) adds +60,000 at Lv1 then +10,000 per level up to +150,000 at Lv10. Buffs Rally Capacity, which is the whole rally's ceiling – not the per-march Squad Capacity. The calc has no separate rally-total variable, so it applies the bonus to the main march slot only (where extra rally room visibly lands); joiner caps are untouched. Same 2h / 20h shape as Fearless Roar.

Valora's Dance of the Hunt and Savage Advantage are not added.They are permanent Master talents, so they are already baked into the capacity the game displays. Adding them again would double-count – enter the cap exactly as the game shows it. (§12.)

How the troops fill the marches:

Your Rally – the march you host – is filled to the Inf/Cav/Arc ratio (§04–§05), capped by what you actually own.

Joiner marches – one per remaining queue – are filled in a fixed order: a small infantry slice first (so each march has front-line soak and the rally clears the 5,000 floor, §13), then archers, then cavalry, then any leftover infantry. Marches with a hero use the with-heroes cap; the rest use the no-hero cap. An alliance join limit, if set, caps every joiner march.

Leader 2nd and 3rd skills – the calculator models only the first skill of each hero. A rally leader's other six expedition skills are real damage sources but their values and effect_op ids are not public.

Joiner-march damage – the damage estimate covers your hosted rally only. Damage from marches you send as a joiner depends on a host you don't control and is not modelled.

Hero skill RNG procs – chance-based skills (Chaos Gambit and the like) are real damage but unmodelled. The estimate is a no-procs baseline.

Truegold combat skills – the archer TrueGold skill (+25% at TG5) is now modelled as an optional toggle, but the infantry and cavalry TrueGold combat skills, and the exact TG3/TG8 thresholds, still have no datamined value table and are not modelled.