Why you can trust it — and where you shouldn't
Escaping the cargo cult
A cargo cult copies the form of something that worked without understanding why it worked. Most KingShot calculators are cargo cults: they reproduce a simplified damage formula from one guide, never cite it, and present every output with the same flat confidence whether the underlying mechanic is documented or guessed.
This project started from the opposite rule — write down every mechanic, mark exactly how it was verified, and never let a guess masquerade as a fact. That discipline is the whole point, and it's visible on every page.
The reliability grades
Every mechanical claim on this site carries one of three grades. They're not decoration — they tell you precisely how much weight a number can bear.
Confirmed
Read or seen directly
Visible in the game UI, stated in patch notes, or readable in the open-source State-of-Survival engine code. The army factor, the per-troop math, the SkillMod formula.
Probable
Reproduced by testing
Not visible in the UI, but reproduced consistently across community testing and multiple matchups. Several op-code assignments and joiner edge cases.
Speculative
Inferred, not yet proven
A logical inference from comparable mechanics, not yet experimentally validated. The fatigue factor and Yang's exact pity rate. These ship as toggles, not assumptions.
The anchor
KingShot's Expedition engine descends from State of Survival, and SoS has an open-source engine implementation plus a published worked example. That example — round zero, infantry on infantry, producing exactly 189 deaths — is the load-bearing validation of the whole simulator.
Confirmed The engine carries the round-zero example as a permanent regression test. Any change to the damage formula, the per-troop computation, the SkillMod resolver or the bonus aggregation has to keep producing 189, or the change is rejected. Hundreds of additional tests pin the rest of the behavior the same way. See the worked example →
Expected vs Monte Carlo
The engine is deterministic in expected mode: every chance-based skill collapses to its average contribution, so the same inputs always give the same answer. That's what makes the Best Counter search fast enough to screen millions of compositions. Monte Carlo mode then re-rolls the survivors with real randomness — default 200 trials — so the final ranking reflects variance, not just the mean. A high-average, swingy rally and a slightly lower-average, steady one are told apart here, not in the average.
Sources
Nothing here is invented. The mechanics trace to a small set of community references, and where they disagree, the disagreement is noted rather than hidden:
- The open-source State-of-Survival engine — the exact algorithm of the original KingsGroup engine.
- Daryl's KingShot guides and working simulator — joiner mechanics, the op-code taxonomy, cross-validation against in-game reports.
- kingshotdata.com — hero stat database, skill descriptions, per-star tables.
- The full SoS combat formula write-up (army factor and fatigue).
- Frakinator's KingShot calculator and community testing from MTwyDev and others.
What it deliberately doesn't do
It doesn't model Conquest (real-time hero duels) or the Bear Hunt damage model — those are different engines. It doesn't simulate garrison leader replacement; you enter the active leader. And it won't hand you a confident number for something nobody has confirmed — it'll give you a toggle and tell you it's a toggle.
Found a value that's wrong?
That's the most useful thing you can bring. The reliability grades exist to be upgraded — come argue it with evidence.