Concepts → Great Weapon Master break-even
Great Weapon Master — at what AC does -5/+10 stop being
worth it?
The single most-asked min-max question in 5e: when does eating a −5 penalty to your attack roll for +10 to damage still come out ahead? It depends entirely on your attack bonus. The break-even AC follows a closed-form rule worth memorising — and the engine computes the underlying damage distributions exactly.
The setup
Great Weapon Master (the 2024 version, "Heavy Weapon Master" in some printings, but the −5/+10 trade is the same in both editions): when you swing a heavy weapon, you can choose to take −5 on the attack roll. If the attack lands, deal +10 damage.
The question is the same one every Fighter, Barbarian, and Paladin asks at character creation: does the trade pay off, or am I just missing more often for nothing?
The math
Two quantities matter:
-
Hit chance. Against AC
Awith attack bonus+B, you hit on a roll ofA − Bor higher on the d20 — so the base hit chance is(21 − A + B) / 20, clamped to the[5%, 95%]band that natural-1 / natural-20 enforces. -
Expected damage on hit. The mean of your
weapon damage roll. For a Greataxe with +3 STR mod, that's
E[1d12+3] = 9.5. For a Greatsword with +3,E[2d6+3] = 10.
Damage-per-round (DPR) without GWM:
DPR_without = hit_chance(A, B) · E[base_damage]
With GWM, hit chance drops by 5 (because the −5 penalty is
equivalent to facing AC A + 5), and damage gains
a flat 10:
DPR_with = hit_chance(A + 5, B) · (E[base_damage] + 10)
Break-even is where these are equal — solve for AC and you get a clean rule.
The break-even table
Numbers below assume a Greataxe (1d12+STR) for the base damage. The break-even AC shifts by ~1 if you use a Greatsword (2d6+STR, mean 10) or a Halberd / Pike (1d10+STR, mean 8.5) — the headline pattern is identical.
Attack bonus (+B) |
STR mod | E[base damage] | Break-even AC | Below this AC, GWM wins |
|---|---|---|---|---|
+5 |
+3 |
9.5 (1d12+3) |
≈ 13 |
AC ≤ 12 — most early-game enemies, plus Knights |
+7 |
+4 |
10.5 (1d12+4) |
≈ 15 |
AC ≤ 14 — covers low-tier monsters and most humanoids |
+9 |
+5 |
11.5 (1d12+5) |
≈ 17 |
AC ≤ 16 — covers everything except plate + shield |
+11 |
+5 + magic weapon +1 |
12.5 (1d12+6) |
≈ 19 |
AC ≤ 18 — covers nearly every enemy in the Monster Manual |
The pattern: break-even AC ≈ attack-bonus + 8 for a Greataxe-class weapon. With a magic weapon at higher tiers, GWM is essentially always on. At low levels with a mediocre attack bonus, it's only worth the trade against lightly-armoured targets.
The damage rolls themselves
The +10 isn't free — it shifts the entire damage distribution
right by 10 points. Click into either panel to see the
per-outcome rationals; tap a bar to read off
P(damage = N) exactly.
1d12+5
- 6 8.33%
- 7 8.33%
- 8 8.33%
- 9 8.33%
- 10 8.33%
- 11 8.33%
- 12 8.33%
- 13 8.33%
- 14 8.33%
- 15 8.33%
- 16 8.33%
- 17 8.33%
1d12+15
- 16 8.33%
- 17 8.33%
- 18 8.33%
- 19 8.33%
- 20 8.33%
- 21 8.33%
- 22 8.33%
- 23 8.33%
- 24 8.33%
- 25 8.33%
- 26 8.33%
- 27 8.33%
Note the shape doesn't change — variance stays at
(12² − 1)/12 = 143/12 for either roll, since the
flat +10 is a constant offset. What changes is the mean and
where the threshold-crossing probabilities land.
When GWM is wrong
Three regimes where the rule breaks down and the table above misleads:
- You have advantage. Advantage on the attack roll changes the hit-chance math non-linearly — see advantage and disadvantage for the curve. Short version: with advantage, the −5 penalty is much less costly because you're rolling twice. GWM becomes correct ~2 AC points higher than the non-advantage table suggests.
- You're using Bless or some other +1d4 to-hit floater. Each +1 to attack expected from Bless is roughly equivalent to facing 1 lower AC. Add Bless's mean (+2.5) to your effective attack bonus when reading the table.
- The target has a kill threshold you're chasing. Mean DPR isn't the question if you need to finish a specific HP threshold in this round — see variance and kill probability. The +10 from GWM helps less than you'd expect when the target is below your single-swing maximum already; it helps more than you'd expect when the target is just above.
Try it yourself
↦ /vs/1d12+3/1d12+13 — base vs GWM, +3 STR ↦ /vs/1d12+5/1d12+15 — base vs GWM, +5 STR ↦ /strike/1d12+5 — Greataxe damage distribution
Drag the HP slider in any /vs view above. Below the GWM mean, the +10 is dominant; well above, the to-hit penalty starts mattering more — same logic as the kill-probability lesson applied per-attack.