Combat Stats: Additive vs Multiplicative
The same 10% is worth 2,010% or 2,200% depending on which bucket it lands in. Which sources multiply, why "by 6%" wording lies, and why the damage formula being a product means balance beats min-maxing Lethality.
The concept in a line — where does the 3000% in the report come from?
Every source feeding your stats is either additive or multiplicative. Most add into one stat, and only a handful multiply on top. Knowing which is which is what lets you decide precisely where to invest and where to send buffs.
⚠️ This video covers Expedition only — rallies, garrisons, city attacks and battlefield combat. The Arena and Exploration are out of scope.
① The same 10% gives different results a +10% source on top of 2,000% Attack
→ A gap of 180 stat points. That is why knowing which sources multiply is decisive. Take every multiplicative source you can, and funnel buffs to the strong players who already hold a lot of them.
🐋 Why whales are stronger than their listed stats suggest
A fully buffed whale on Expedition gets widget 15% + pet 10% + buffs 20% = 45% multiplied in → 10 Lethality behaves like 14.5 (nearly +50%). On the battlefield, the 50% turret multiplies in on top. That is why identical stat lines can mean double the real damage or tanking.
② Every additive source they all add together into one stat
| Source | Applies to | Notes |
|---|---|---|
| Hero stars | Attack / Defence | Newer generations carry bigger stats, which is why new heroes are strong. A large source of Attack and Defence |
| Hero Gear | All | Enhancement and Mastery, all of it |
| Expedition stats (Lethality and the rest) | All | Rallies, garrisons, city attacks, the battlefield |
| Class widgets (Infantry / Marksman Health and Lethality) | Per-class Health and Lethality | The additive part of a widget (the multiplicative part is in ③ below) |
| Natalia and Jeronimo's skills | All | Written as "by 6%", which reads multiplicative, but measured to be additive |
| Chief Gear · Chief Charms | Attack, Defence / Health, Lethality | 100% additive, no dispute |
| Research (the War Academy) | All | All of it, both the all-class and single-class lines |
| VIP (from level 9 up) | Attack / Defence | VIP9 10% → VIP10 12% … |
| Alliance tech, facilities and territory | All | Including the 10% Attack/Defence inside the flag |
| Islands (trees and buildings) | All · Attack / Defence | Every source here is additive |
| Skins (portraits, teleports, name tags and so on) | All | The amounts are tiny — do not buy them for the stats |
| State officer titles, Frostdragon Tyrant titles, President buffs | Lethality · Attack · all | Sunfire and Tyrant stacks also add with each other |
| Domain (globe skins 1–6) | Attack, Defence, training and gathering speed | Higher levels give a bigger % |
| Experts (Romulus and the rest) | All | The % wording confused people, but testing plus customer support confirmed it is additive |
| Bear Trap arrows (≤25%) · Foundry ammunition (15%) | Temporary | The values are additive |
⚠️ The wording trap — things that say "by 6%" but add
The game writes "by ~%", which reads as multiplication, but these are additive in practice:
- Natalia (Ursus Strength) and Jeronimo (Sword of Jeronimo / Natural Leader) — confirmed additive by repeated testing.
- The Expert Romulus — the old % wording had everyone convinced it multiplied, but testing plus customer support confirmed additive (and the wording was later fixed). 10% on 3,000% is just 3,010%. Romulus is for rally and troop capacity, not stats.
The lesson: never assume something multiplies just because it is written as a %. Test it.
③ Every multiplicative source multiplied on top of the additive bucket → total damage explodes
| Source | Value | Applies to | Conditions |
|---|---|---|---|
| Hero widgets (the stat-% type of skill) | Lv2 5% → Lv10 15% | Attack or Defence | Conditional — split between attacking-rally and garrison versions. Does not fire on a solo march |
| Pet skills (triggered, temporary) | 8–10% each | Attack / Defence / Health / Lethality | Enemy-debuff pets multiply too |
| City buffs | 10% / 20% | All | Sunfire and events |
| Frostdragon Tyrant turrets | 50% | Attack/Defence or Health/Lethality | During the battlefield only, 30 minutes per hour |
| Tyrant crystals, patrols and mine buffs | Varies | All | Battlefield only · additive with the turrets when both are up |
| Frostfire Mine, left side Lv5 | Lethality +50% | Lethality | Battlefield only, all classes |
| Expert Arena boosts (left = escorts, right = heroes) | — | Attack · Health | The only multiplicative Expert effect (99% confidence) · Arena |
| ⭐ Hero skills (Patrick, Lynn, Bradley) | e.g. Health +25% | — | Not shown in the report; the same type stacks additively |
🔧 Why widget Lv10 is essential
The multiplicative part of a widget scales with its level, Lv2 5% → Lv10 15% (Attack or Defence). Whales multiply that against 2,000–3,000% Defence for another 300–450% of defence. Because it is conditional, only the defensive widget fires on a garrison and only the offensive one on an attacking rally — which is why attack and defence setups use different heroes and widgets.
🐾 Why maxed pets matter so much for a rally leader
A pet's levels, books, serums, potions and refinement are additive; only the triggered skill multiplies. Which is why 4–6 pets can beat a stronger rally from someone with none. The late-game debuff pets (the Polar ones and so on) multiply too. This is the core reason leaders need to rotate every 5–6 hours.
🎁 Send buffs to the strong
Ordinary buffs (10/20%) multiply, so applying them to low stats gains little → give them to whoever has the deployment size, Health and Lethality. Conversely officer, President and Championship buffs are additive, so adding 10% to 2,000% Lethality is nearly meaningless — no need to fear an attack under a Sunfire buff (though they are useful for beasts and Championship registration).
🏰 The Tyrant battlefield is a multiplication party (but nothing burns)
With the turret's 50% multiplier (30 minutes an hour), an alliance stacking defence and health with defensive heroes and widgets gets buffs 20 + widget 15 + pets 10 + turret 50 = 95% → nearly double (200% health becomes 400%). But those multipliers do not work off the battlefield → cities do not burn. Controlling the middle is what matters. The Frostfire Mine works the same way — if the enemy turns on their +50% Lethality, wait out the cooldown (1–2 minutes) and then march.
④ How to spot multipliers in the report the exclamation mark — except hero skills, which never show
| Where in the report | What it is |
|---|---|
| Top of the exclamation mark (!) | buffs Both sides' Attack, Defence and similar buffs |
| Below that | widgets Defence = 15 Defence and 15 Lethality / attack = 15 Attack (45% when stacked) |
| Below that | pets If they had theirs on and you did not, this is where the result gap came from |
| Nowhere at all | hero skills Patrick's +25% Health and the like never appear in the report → only confirmable through community testing |
※ During a live attack you can see the skills the top 4 joiners contributed (3 × Jessie + 1 × Renee, say). When defending a garrison, tap the rally flag to check the shared skills and kick anyone running a hero other than the one you asked for, so they can be replaced. Verify against the report by comparing screenshots of the top 4 joiners. → For the detail, this is the same phenomenon as the "hidden skill damage" in the normal attack vs skill damage guide.
⑤ The damage formula video ② · why balance is the answer
❌ "Lethality beats Attack" is only half right
Because the formula multiplies, Attack and Lethality are symmetric. For the same total of 300, balance beats skew → Lethality is not inherently superior; raising whichever is lower toward balance is what grows the product. The old line that "Attack and Defence are bad, only Lethality and Health matter" rested on a wrong premise (Ton included — he deleted his earlier, error-filled video over it).
| Attack | Lethality | Damage (÷100) |
|---|---|---|
| 200 | 100 | 200 |
| 100 | 200 | 200 |
| 150 | 150 | 225 |
📉 Feed the lower stat first — +10 on Attack 1,000 / Lethality 600
Adding the 10 to the lower one (Lethality) wins: 1010×600/100 = 6,060 against 1000×610/100 = 6,100 → 41 points more. A +10 is proportionally larger against a smaller number. Default rule: raise whichever of your own stats is behind (reverse it if your Health is far above your Attack and Defence). Apply this when packs, event rewards and research start to slow down.
Note this is not because Attack and Defence are bad — they climb easily on their own from heroes (even a rally leader sits at 3,000% Attack/Defence against 2,000% Lethality), so when you get a choice you push the Lethality and Health that do not climb by themselves, to restore the balance.
💰 Value — do not throw away cheap, easy Attack and Defence
Since they are symmetric, price and difficulty count too. If 10% Defence costs $1 and 10% Lethality costs $10, the $1 of Defence may well be the better buy (and if you can afford both, $11 gets you both). Extreme min-maxing — huge Lethality against very low Attack — does not work, because there is nothing for the Lethality to multiply against.
🔧 Chief Gear, reassessed
Having understood the formula, Ton corrected his earlier advice: the Chief Gear he used to call "useless" is not as bad as he thought — expensive per %, but permanent (hero stats are temporary) and it adds a little deployment capacity. The priority order still stands, though: Charms > Chief Gear, and Hero Gear > Chief Gear. If you are at the very bottom of the investment curve, catching up via the $5 weekly pack is good value, because your Attack and Defence get multiplied against the Lethality and Health you have already built. Research is also at its best value on the first level of each line — spread a little across many lines.
⑥ Stat priority where it goes, by class
| Class | Priority stats | Why / in practice |
|---|---|---|
| Infantry | Defence and Health | Tanks. You can relatively skip Lethality and Attack |
| Marksman | Attack and Lethality (plus Defence and Health) | Your damage, but they melt without Lancers and Infantry → do not ignore Defence and Health |
| Lancer | Defence and Health (plus Attack and Lethality) | Defence and Health matter most, but you want both |
Push each class's good axis to Lv4–5 first and defer the bad one (Infantry Lethality, Marksman Defence and so on). Specialising your leader — 60/40 (Greg) or 50/50 (Mia) — and splitting your stat investment along the same lines is efficient: a Gen3 Mia 50/50 scales every class (no Lancers needed), while a Gen3 Greg 60/40 tanks with Lancers too.
⑦ Verification the additive model matches the report exactly, 0.0% difference (2025-07-07)
📐 Profile + heroes + Jeronimo and Natalia = the report, with zero error
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 simply add. Individual sources were confirmed with isolation tests as well.
| Class | Stat | Profile | +heroes | +Jer. & Nat. | =total | Report | Diff |
|---|---|---|---|---|---|---|---|
| Infantry | Attack | 324.9% | 559.0% | 4.0% | 1212.7% | 1212.7% | 0.0% |
| Infantry | Defence | 313.9% | 589.0% | 4.0% | 1273.4% | 1273.4% | 0.0% |
| Infantry | Lethality | 420.5% | 260.0% | 6.0% | 747.2% | 747.2% | 0.0% |
| Infantry | Health | 433.6% | 412.7% | 6.0% | 929.0% | 929.0% | 0.0% |
| Lancer | Attack | 228.2% | 540.4% | 4.0% | 1097.4% | 1097.4% | 0.0% |
| Marksman | Attack | 250.8% | 560.4% | 4.0% | 1140.0% | 1140.0% | 0.0% |
An extract — Lancer and Marksman Defence, Lethality and Health all matched at 0.0% too. 💡 A note from the sheet: these days Ton recommends raising Marksman Health over Lancer Health.
✅ In one line
Most sources are additive and add into one stat; only widgets (at Lv10), pet skills, city buffs, the Frostdragon Tyrant turrets, the Frostfire Mine and hero skills multiply, and that is what makes total damage explode. Take every multiplier you can and funnel buffs to the strong. Never assume "by %" means multiplication — test it (Natalia, Jeronimo and Romulus are all additive). On stats, do not min-max Lethality and Health alone: lay down Attack and Defence broadly, then specialise by class (Infantry = Defence/Health, Marksman = Attack/Lethality, Lancer = Defence/Health).
Widget %, pet % and the 50% turret are Ton's figures · the damage formula is the community standard derived from the verified State of Survival formula · hero and generation matching per the vault verification table (2026-07-24) · the full version of the appendix to the normal attack vs skill damage guide.
Naming corrections: subtitles' JC → Jessie, Geronimo → Jeronimo. ※ "Sword of Jeronimo" does not appear in the wiki's skill list (Swordmentor, Expert Swordsmanship, Natural Leader) and needs rechecking against the original.