All guides

Normal Attack vs Skill Damage

Total damage is normal attack plus skill damage, and you cannot value a joiner without knowing which one a hero pushes. Reina raises only the former, which is why she underperforms at the Bear Trap, and Wu Ming only beats Patrick against skill-heavy opponents.

공략노트·TontonYouTube basics video, plus Ton's public additive/multiplicative stats sheet. The hidden skill damage list is a community work in progress and only runs to Gen 11
① Normal AttackEvery point of damage that is not skill damage — your troops' basic hits. This is what Reina raises.
② Skill DamageDamage a skill deals directly on trigger. Added in bulk from Gen 5–6. This is what Wu Ming cuts (−30%).

The concept in a line

Nearly all combat damage in WOS is either ① normal attack or ② skill damage. Your total damage = normal attack + skill damage. (A very small number of exceptions exist.)

Hero and troop buffs raise one of the two, or both. You cannot work out a joiner's value without knowing which of the two a given hero pushes, which is why this document is the prerequisite for any joiner or rally guide.

① Why the distinction matters three representative heroes

HeroWhat they doThe catch / conditions
ReinaRaises normal attack onlyPut her in a rally full of skill damage (the Bear Trap, say) and she only inflates a small slice of the total → what you feel is 15–20%, not 30%
Wu MingAgainst Infantry: normal attack −25% / skill damage −30%Only strong when the enemy front line carries a lot of skill damage. If it does not, he is just a flat 25% reduction against Infantry
Patrick+25% Health to every classThe usual comparison against Wu Ming. Wu Ming only wins under the "enemy is skill-damage heavy" condition (Patrick covers Lancers and Marksmen too)

② Wu Ming vs Patrick when is it Wu Ming?

⚖️ How to decide

Wu Ming's reduction applies to Infantry only and cuts harder on the skill-damage side (−30%). Patrick gives every class +25% Health (≈ 25% less damage taken).

  • Enemy front line is light on skill damage → Wu Ming ≈ a 25% reduction for Infantry alone → Patrick is better (he covers Lancers and Marksmen).
  • Enemy front line is heavy on skill damage → Wu Ming's reduction averages up toward ~27.5%Wu Ming is better.
  • Lancers are rarely targeted by skills → Wu Ming's value rises when you are not running Marksmen and Infantry survival is what matters.

🛡️ In practice — 2 × Wu Ming with a 60/40

Wu Ming (the skill-damage reduction) arrived as the counter to the flood of skill damage added in Gen 5–6 — the game's usual rhythm of "a powerful tool, then a counter to it next generation". If your opponent is Gen 6 and full of Waynes and you are struggling → 2 × Wu Ming with a 60/40 can be very strong. He works especially well on a defensive (garrison) 60/40.

③ Finding skill damage in the report

🔎 It is not on the first page → open "Battle Detail"

The report's first page only shows troop counts. Go into Battle Detail at the bottom left and the skills are listed in a fixed order: the top 3 hero skills (Infantry hero 3 → Lancer hero 3 → Marksman hero 3), then the troop skills below them.

What the report showsWhat it means
A kill count attachedDefinitely skill damage — the basic test (Crystal Gunpowder, for instance, racks up kills)
Only raises Attack or LethalityNot direct damage but a buff (it raises normal attack or skill damage)
Skill damage with no visible killsHidden skill damage — only confirmable by testing (Rufus S2, for example). See list ④ below

Infantry do not produce skill damage (Infantry heroes such as Eleonora deal in damage reduction and buffs). ※ The S numbers count the three Expedition skills on the right of the hero's skill panel, right to left (S1 = rightmost). Rallies and joiners always use the Expedition skills.

④ Heroes with hidden skill damage WIP · Gen 1–11 · not shown in game

⚠️ A work-in-progress list

Skills that deal real skill damage without the game showing it. This is a community-crowdsourced list that Ton himself calls a "first version, incomplete", and it only runs to Gen 11. (Skills that show kills are excluded — you can see those in the report directly.) Some slot numbers may be off, given the auto-generated subtitles they came from.

GenHeroClassSlotNotes
G1MollyLancerS2
G3MiaLancerS2
G4ReinaLancerS3⚠️ Her headline skill (the joiner S1) raises normal attack, but S3 is skill damage
G5GwenMarksmanS2A lot of skill damage
G5HectorInfantryS3
G5NorahLancerS2
G6WayneMarksmanseveralThree of his skills are skill damage → which is exactly why Wu Ming's reduction bites so hard against him
G7GordonLancerS1May work like Molly's S1
G8SonyaLancer
G9FredLancerS3
G11RufusMarksmanS2Confirmed from the report

🔀 The exception — Renee (G6 Lancer)

Renee's damage is neither normal attack nor skill damage. That makes her very good but a special case that scales poorly with other buffs and debuffs. Data past Gen 11, and especially Gen 14+, is thin because there are few testers → new skills get added to this table as they are found.

⑤ What it means in practice judging joiners

❌ Why Reina is a poor fit for the Bear Trap

At the Bear, the leader's team runs Mia and most of the Marksmen deal skill damage (Hector, Gwen and Wayne included) → a large share of the rally's total is skill damage. Reina raises normal attack only, so she gives 30% of that smaller slice → 15–20% in practice. A guaranteed 25% joiner (Jessie, Jasser, Seo-yoon, Bradley, Magnus, Blanchette) is better. By Gen 5–6, Reina is not a suitable Bear joiner.

❌ An example of using Wu Ming wrongly

Running Eleonora + Lloyd + Rufus as leaders on a Marksman-heavy build and adding Wu Ming is a mistake:

  • The only confirmed skill damage there is effectively Rufus S2 → there is nothing for Wu Ming's −30% to bite on.
  • Wu Ming's −25% is Infantry-only → and if you are fielding Marksmen, they get sniped by Lancers and skills and simply melt.
  • Eleonora and Lloyd deal no skill damage (confirmed by testing). Lloyd's "enemy Lethality down + Attack +150%" looks like skill damage but is not.

⑥ How any of this gets verified the controlled T5 test

🧪 The game only tells you how many troops were injured

The game never shows a damage number, only injury and kill counts. So you work backwards with a controlled experiment.

  1. Baseline — collide T5 (FC5) troops with no skills. Strip out heroes, joiners and buffs and there is no randomness, so the result is identical every time.
  2. One variable — add a single Lloyd (the opponent stays the control) and run 3,000 T5 vs 3,000 T5 repeatedly, say 10 times → take the average injury count.
  3. Add the suspect hero — add Reina and see whether the injury count (total damage) actually rises → from which you deduce whether it is normal attack or skill damage.
  4. Isolate the verdict — "does Wu Ming reduce the damage? yes" / "does Reina increase it? yes → skill damage".

This is the accumulated work of the community over the last two years. Verification is solid through Gen 11–12; beyond that there are not enough testers. A result only counts as "confirmed" when it reproduces in real rallies, not just in theory.

⑦ Appendix · additive vs multiplicative stats Ton's public sheet

🔗 The material that follows directly from hidden skill damage

The sheet in one line: hero skills (Patrick, Lynn, Bradley and their kind) do not appear in the report, and stacking the same type adds while different types multiply. This table is the underlying basis for the joiner-bucket maths in the participant hero guide.

The big picture: almost every permanent growth source feeding your stats lands in a single additive bucket, and only a handful of multiplicative sources are applied on top.

Additive (everything here sums together)Applies toNotes
Hero stats (stars, shards)Attack / DefencePermanent while they are in your march, temporary otherwise
Hero Gear / hero widgets (Health and Lethality %)All · Health / Lethality
Jeronimo (Natural Leader) · Natalia (Ursus Strength)AllAdditive in this context
Chief Gear score / Chief CharmsAttack, Defence / Health, Lethality
Research · VIP 9–12 · alliance tech · facilitiesAll
Skins · islands · territory · pet level and refinementAll · Attack / Defence
State officer and Frostdragon Tyrant titles · SvS President Lethality · Domain (gauntlets 1–6)Lethality / Attack · all
Experts (Romulus · Gareth S2 and S3)All
Bear Trap arrows (up to 25%) · Foundry ammunition (15%)TemporaryAmmunition applies while you hold the building
Multiplicative (these multiply each other → total damage explodes)Applies toNotes
Hero widgets (the stat-raising skill-% type)AllConditional — split between garrison and attacking-rally versions (read the description)
Pet skill % (2h) · city bonuses (10/20%)All
Frostdragon Tyrant turret · Crystal skillsAll · Attack / DefenceIf both are active, those two are additive with each other
Hero skills: Health and Defence (e.g. Patrick)Not shown in the report; additive when stacking the same type
Hero skills: Attack up and damage (e.g. Lynn, Bradley)Not shown in the report; additive when stacking the same type
Gareth S4 (enemy Lethality down) · Holger in the Arena · Frostfire Mine left S Lv5 (Lethality +50%)Lethality · Attack & Health
❓ Troop skills (Master Brawler · Bands of Steel · Body of Light)???Unresolved even in the sheet

📐 Verification — the additive model matches the report exactly, 0.0% difference (2025-07-07)

Ton summed profile (military) + heroes + Jeronimo and Natalia and compared it against the actual report → 0.0% difference across every class and every stat. Empirical proof that additive sources really do just add up.

ClassStatProfile+heroes+Jer. & Nat.=totalReportDiff
InfantryAttack324.9%559.0%4.0%1212.7%1212.7%0.0%
InfantryLethality420.5%260.0%6.0%747.2%747.2%0.0%
InfantryHealth433.6%412.7%6.0%929.0%929.0%0.0%
LancerAttack228.2%540.4%4.0%1097.4%1097.4%0.0%
MarksmanAttack250.8%560.4%4.0%1140.0%1140.0%0.0%

An extract — Lancer and Marksman Defence, Lethality and Health all matched at 0.0% across the board too. 💡 A note from the sheet: these days Ton recommends raising Marksman Health over Lancer Health.

✅ In one line

Total damage = normal attack + skill damage. In the report's Battle Detail, a kill count means skill damage; for the hidden ones, use the WIP list above. Reina raises normal attack only, so she is a poor fit for the skill-damage-heavy Bear Trap. Wu Ming is −25%/−30% against Infantry, so he is only strong when the enemy is skill-damage heavy (a 60/40 against Waynes) — otherwise take Patrick. Verify anything with a controlled T5 test, comparing injury counts.

Source: TontonYouTube — Normal Attacks & Skill Damage · appendix: Ton's public additive/multiplicative stats sheet
Wu Ming's 25%/30% figures cross-checked against Heaven Guardian (Gen6) · hero and generation matching per the vault verification table (2026-07-24).
The hidden skill damage list is Ton's work-in-progress and is still being updated.
Naming corrections: subtitles' JC → Jessie, Raina → Reina, Rene → Renee, Wuming → Wu Ming, Sonia → Sonya, Nora → Norah.