How Kill Team (2024) balances weapons against armor, and what we can take — 2026-09-05
Why we looked: the heavy investigation found that the light roster has no anti-armor answer except a Cannon arm of its own, and the design lead wants glances to stay at half damage so armor is a wall. Kill Team's third edition (2024) is the closest published game to our problem: a skirmish with a handful of operatives per side, dice-for-dice resolution, saves that are meant to matter, and an active balance program. We cannot copy it (no hit locations, no roster building, no points) but its dice math and its process are both worth reading.
Sources: the core rules and datasheets as transcribed on Wahapedia (Kill Team 3), the ktcalc weapon-rules reference, and Warhammer Community's balance dataslate articles. Numbers below were recomputed by enumeration (kt-math.mjs, in the session scratch) rather than copied.
1. The core loop, in our vocabulary
Shooting. The attacker rolls Atk d6 (almost always 4, heavy weapons 5). Each die at or above the weapon's Hit stat is a normal success; a 6 is a critical success. The defender rolls three defence dice regardless of how many hits came in; each die at or above the operative's Save is a normal save, a 6 a critical save. Saves block hits: a normal save blocks a normal hit, a critical save blocks either, and two normal saves together block one critical hit. Every unblocked normal hit does the weapon's Normal Dmg, every unblocked crit its Critical Dmg (a weapon reads 3/4). No damage dice: damage is a stat, variance comes from hits and saves.
Cover lets the defender retain one of its three dice as a normal save without rolling it. Obscured (heavy terrain between) makes the attacker discard one success and downgrades all its crits to normals. Nothing else modifies the roll: no to-hit modifiers, no armor subtraction.
Melee is symmetric: both roll their weapon's Atk dice, then alternate resolving successes as a strike (deal damage) or a block (cancel one of the opponent's successes; crits block anything, normals block normals only).
Stats. APL 2–3 actions per activation, Move 3–7", Save 2+ to 6+, Wounds 7–18. Elite operatives get Save 2+/3+, 14–18 Wounds and APL 3; hordes get Save 5+, 7 Wounds, APL 2 and twice the bodies.
2. The weapon vocabulary
Every weapon is Atk / Hit / Dmg N/C plus zero to three keywords. The keywords are the whole design space, and each one moves exactly one axis:
| Keyword | Rule | Axis it moves |
|---|---|---|
| Accurate x | retain x dice as normal successes without rolling | hit floor |
| Balanced / Ceaseless / Relentless | re-roll one die / all dice of one value / any dice | hit consistency |
| Lethal x+ | crits on x+ instead of 6 | crit rate |
| Rending | with any crit, one normal becomes a crit | crit rate |
| Severe | with no crit, one normal becomes a crit | crit floor |
| Punishing | with any crit, one fail becomes a normal | volume |
| Piercing x | defender collects x fewer defence dice | armor denial |
| Piercing Crits x | Piercing only if a crit was retained | conditional armor denial |
| Devastating x | each retained crit does x damage immediately, before saves | unblockable damage |
| Saturate | defender gets no cover save | cover denial |
| Seek / Seek Light | target cannot use (light) terrain for cover when selected | cover denial |
| Brutal (melee) | opponent can only block with crits | parry denial |
| Shock (melee) | first crit strike also discards an opponent's normal | tempo |
| Stun | with any crit, target loses 1 APL next activation | tempo |
| Heavy | cannot move and shoot (or Dash only) | positioning cost |
| Hot | after firing roll a d6; below the Hit stat, take twice that in damage | self-damage for power (our heat) |
| Range x | targets within x only (pistols 6–8") | reach |
| Blast x / Torrent x | secondary targets within x | area |
| Limited x | x uses per battle | scarcity |
| Silent | may shoot while Concealed | stealth |
The mapping to our system is close: Piercing is our Pen, Devastating is our "ignores Armor", Lethal is our crit margin, Heavy is our Heavy tag, Hot is our weapon Heat, Range is our range bands, Limited is a knob we do not have yet.
3. Reference profiles
| Operative | APL | Save | Wounds |
|---|---|---|---|
| Death Korps trooper | 2 | 5+ | 7 |
| Kasrkin-class trooper (typical mid) | 2 | 4+ | 9 |
| Intercessor (Space Marine) | 3 | 3+ | 14 |
| Custodian Guard | 3 | 2+ | 18 |
| Weapon | Atk | Hit | Dmg | Keywords |
|---|---|---|---|---|
| Lasgun | 4 | 4+ | 2/3 | |
| Boltgun | 4 | 3+ | 3/4 | |
| Bolt rifle | 4 | 3+ | 3/4 | Piercing Crits 1 |
| Heavy bolter (focused) | 5 | 3+ | 4/5 | Piercing Crits 1 |
| Krak grenade / launcher | 4 | 3+ | 4/5 | Piercing 1 |
| Plasma gun (standard) | 4 | 3+ | 3/5 | Piercing 1 |
| Plasma gun (supercharge) | 4 | 3+ | 4/5 | Hot, Lethal 5+, Piercing 1 |
| Meltagun (Krieg) | 4 | 4+ | 6/3 | Range 6", Devastating 4, Piercing 2 |
| Bolt sniper (mortis) | 4 | 2+ | 3/3 | Devastating 3, Piercing 1, Heavy, Silent |
Note the shape of the anti-armor weapons: the meltagun's normal damage is higher than its crit damage because the crit's value is the unblockable Devastating 4. Plasma buys Piercing with a supercharge mode that hurts the user. Every team, including the 14-body conscript horde, gets two or three of these on gunner slots.
4. The math: how much armor actually stops
Expected damage per Shoot action, defender blocking optimally, target in the open / in cover:
| Weapon | Krieg 5+ W7 | Save 4+ W9 | Marine 3+ W14 | Custodian 2+ W18 |
|---|---|---|---|---|
| Lasgun 4A 4+ 2/3 | 2.8 / 2.0 | 2.2 / 1.6 | 1.6 / 1.3 | 1.0 / 0.9 |
| Boltgun 4A 3+ 3/4 | 5.7 / 4.1 | 4.4 / 3.3 | 3.3 / 2.7 | 2.3 / 2.1 |
| Bolt rifle, Piercing Crits 1 | 6.2 / 4.6 | 5.2 / 4.0 | 4.2 / 3.5 | 3.3 / 3.0 |
| Heavy bolter 5A 4/5, Piercing Crits 1 | 10.9 / 8.5 | 9.4 / 7.7 | 8.0 / 7.0 | 6.7 / 6.2 |
| Krak 4/5, Piercing 1 | 8.7 / 6.4 | 7.5 / 5.8 | 6.3 / 5.3 | 5.2 / 4.8 |
| Plasma 3/5, Piercing 1 | 7.1 / 5.5 | 6.2 / 5.0 | 5.3 / 4.6 | 4.4 / 4.1 |
| Plasma supercharge 4/5, Lethal 5+, Piercing 1 | 9.4 / 7.6 | 8.4 / 7.1 | 7.3 / 6.6 | 6.3 / 6.1 |
| Meltagun 6/3, Devastating 4, Piercing 2 | 11.0 / 7.9 | 10.2 / 7.9 | 9.4 / 7.9 | 8.6 / 7.9 |
Shoot actions to kill in the open:
| Weapon | Krieg | Save 4+ | Marine | Custodian |
|---|---|---|---|---|
| Lasgun | 2.5 | 4.2 | 9.0 | 17.7 |
| Boltgun | 1.2 | 2.0 | 4.3 | 7.9 |
| Bolt rifle | 1.1 | 1.7 | 3.3 | 5.5 |
| Plasma | 1.0 | 1.5 | 2.6 | 4.1 |
| Meltagun | 0.6 | 0.9 | 1.5 | 2.1 |
Damage retained against the best save (Custodian 2+) as a share of damage against the worst (Krieg 5+): lasgun 36%, boltgun 40%, bolt rifle 53%, krak and plasma 60–62%, supercharged plasma 68%, meltagun 78%.
Three things stand out.
- The wall is real and steep. A lasgun keeps 36% of its damage into a 2+ save; the horde needs eighteen shooting actions to drop one Custodian. Kill Team does not soften this. It gives the horde a meltagun.
- Saves are a fixed budget, so volume beats armor even without Piercing. Three defence dice can block at most three normal hits. The heavy bolter's fifth die and 4/5 damage keep 61% into a 2+ save with only a conditional Piercing 1, because the fourth and fifth hits have nothing left to stop them.
- Crits are the leak. A crit needs a crit save or two normal saves, so even Save 2+ lets roughly a sixth of hits through at full value. Piercing and Devastating widen that leak; Lethal makes it more frequent. No weapon ever ignores a save completely and no save ever stops everything.
For comparison, our F5 with glance at half: against Heavy core Armor 8 the rifle, autocannon and pulse arm land 58% of their dice per hit (a natural 6 always penetrates), the cannon 75%; against Medium Armor 6 the rifle lands 75%, the autocannon 67%; against Light Armor 4 the rifle 92%. So our worst case (58%) is gentler than Kill Team's lasgun (36%), and our rifle against a heavy already behaves like their plasma gun against a Custodian (60%). The difference is not the fraction. It is that every Kill Team roster has a plasma or melta on a slot, and ours has one anti-armor weapon in the whole catalog, on a 13-weight heavy arm.
5. How they balance power
- No points. A kill team is a fixed list (a leader plus five to thirteen operatives, one of each specialist). Power is balanced per team, not per model or per weapon. The unit of balance is the roster.
- Elite versus horde is balanced with four dials at once: operative count (6 marines vs 14 Krieg), APL (3 vs 2), Wounds (14 vs 7), Save (3+ vs 5+), plus team rules (marines may shoot twice per activation). Scoring rewards board presence, which the horde has, against killing power, which the elite has.
- Weapons are balanced by profile shape, not price. A weapon's cost is the slot it occupies and the keyword tax it pays: Heavy for the sniper, Hot for supercharge, Range 6" for the meltagun, Limited for grenades. Anti-armor is always paired with a drawback.
- The process is quarterly and data-driven. The studio publishes a Balance Dataslate each season. Their stated target is a 45% to 55% win rate per team, read from events of 20 or more players, weighted by pick rate (teams with low pick rates get less trust) and podium finishes (a team can have a modest win rate but a high skill ceiling). Changes are deliberately small and targeted: a Dmg point, a keyword added or removed, an equipment cost, a ploy, a Wounds or APL point. They also hold changes back after a big release "to see how things settle with all the new variables".
6. What transfers to mecha-skirmish
| Theirs | Ours today | Worth taking |
|---|---|---|
| Save = 3 dice per attack, a fixed budget | Every shot rolls its own penetration, so armor scales with volume | Consider a volume leak: shots of one Burst against the same part share one penetration die, each extra hit adds +1 Pen. Makes the autocannon an anti-armor tool by volume, as the heavy bolter is |
| Piercing on gunner weapons in every team | Pen 4 on one heavy arm | Put an anti-armor option within reach of every runner: ammo loads (section 7) or a light high-Pen weapon |
| Devastating: damage before saves | Falls and recoil ignore Armor; no weapon does | A Devastating tag for a melta-class weapon: crit damage ignores Armor. Cheap to add to F5 |
| Lethal 5+ | Crit margin Armor + 4 | Per-weapon crit margin (Lethal) is the same lever with a finer grain |
| Hot / supercharge | Weapon Heat, Overcharge (G3) is an extra action | Energy weapons get an overdrive mode: more dice or Pen for extra Heat. Pairs with ammo for ballistics |
| Cover = one guaranteed save; Obscured downgrades crits | Cover = 1–2 difficulty dice | Hard cover downgrading crits to full instead of a die is worth a sweep: it caps the ceiling rather than lowering the floor |
| Range is mostly unlimited; Conceal orders and heavy terrain limit shooting | Range bands decide open ground (T4) | A Conceal state (cannot be targeted behind hard cover unless you fired) is their answer to range dominance and would address T4 without touching the cannon |
| Heavy (Dash only) | Heavy tag, 1[−] after moving | Same idea, ours is milder |
| 45–55% band, judged on real play, small changes, seasonal cadence | Register with win rates from the bot | Adopt the band per roster archetype at best-response aggression (TASK-021); keep changes to one number or one tag per row |
7. The ammunition idea, with numbers
The design lead's proposal: conventional guns choose ammunition, energy weapons always fire energy. This is Kill Team's gunner-slot answer expressed through our roster-building axis (weight), so it fits the pillars: it is a visible choice on the sprue and a visible die on the table.
Shape. Ballistic weapons (Rifle, Autocannon, Cannon, Missile rack) have a Standard load. A roster may buy one alternate load per weapon for weight; switching loads is free at the moment of Fire. Energy weapons (Pulse, and any laser or plasma to come) have no loads but get an Overdrive mode for Heat, the way Kill Team's plasma supercharges.
| Load | Effect (proposal) | Kill Team analog | Weight |
|---|---|---|---|
| Standard | as printed | 0 | |
| Armor-piercing | Pen +2, damage one step down (3d6 → 2d6+2, 2d6 → 1d6+3, 1d6 → 1d3+1) | Piercing | +2 |
| High-explosive | damage +1d6 on Full or Crit, Pen −2, a glance does 1 | Devastating-ish | +2 |
| Incendiary | on hit 1d3 Heat to the target, Pen −1 | Hot in reverse | +1 |
| Flechette | +1 shot, Pen −2, damage one step down | Torrent-ish, anti-light | +1 |
Why weight and not heat. Weight is the roster decision the game is built around, and the Team Cap is what stops "everyone carries AP". A heavy roster spending 143 of 145 cannot add loads; a light roster at 129 can add four. That is the counter-pressure the heavy roster currently faces nowhere.
What it does in the simulator today (rifle Pen 2 → 4 as a stand-in for an AP load with no damage step, standard rosters on Quarry, 300 games; duels 600 games):
| Light roster win | Courier vs Bastion | Foreman (autocannon Pen 1 → 3) vs Bastion | |
|---|---|---|---|
| Before | 21% | 39% | 35% |
| After | 22% | 47% | 36% |
AP fixes the unit fight (a Courier with AP is close to even with a Bastion) and does nothing for the roster game, because the roster game is decided by cannon range before anyone shoots the Bastion. Ammunition is the anti-armor answer; cannon range 12 → 10 (TASK-022) is the other half, and they should land together.
Decisions this needs before it is a rule. One load per weapon or a magazine of several; whether special loads are Limited (Kill Team's grenades are, and it would make the switch a moment); whether the AI should be taught to pick loads before we trust the numbers; and whether the Pulse arm's Overdrive is the same knob as G3 Overcharge or a separate one. Tracked as IDEA-005.
8. Recommendations
- Keep glance at half. Kill Team's wall is steeper than ours and it works because anti-armor is available, not because the wall is soft.
- Add anti-armor access, not armor reduction: ammunition loads for ballistics (section 7), or failing that a mid-weight high-Pen low-damage arm for Kessler and Meridian.
- Ship cannon range 12 → 10 with it (TASK-022) and re-baseline both at best-response aggression.
- Try the two cheap structural ideas in a sweep each: Burst volume leak against a single part, and hard cover downgrading crits.
- Adopt their balance discipline as ours: a 45–55% band per roster archetype, one change per register row, re-baseline after each.
Sources: Wahapedia Kill Team 3 core rules and the Angel of Death, Death Korps and Talons of the Emperor team pages; ktcalc weapon rules reference; Warhammer Community, "This season's Kill Team balance dataslate is here" (designers' 45–55% target and method); Goonhammer and Tabletop Battles dataslate coverage for the kinds of changes made.