Diceplots

Concepts → Finishing Strahd

Finishing Strahd — how many rounds does it really take to drop 144 HP?

The canonical Curse of Strahd boss fight question. Strahd has 144 HP at his standard CR-15 stat block. Three or four PCs hitting him for 30-40 damage per round each — should the fight last 2 rounds or 5? The renewal-theorem result from the strikes-to-kill pillar gives a clean closed-form answer, and the variance around it is smaller than table-talk suggests.

The setup

Strahd, the Vampire Lord of Barovia, ships at HP 144 in the core stat block. A typical level-10 party that's reached him has roughly:

  • A Greatsword Fighter (Champion) with GWM at +9 to-hit, dealing 2d6+5 base or 2d6+15 with GWM. Two attacks.
  • A Sorcadin / Paladin with Smite. 1d8+5 + 2d8 Divine Smite at 2nd-level slot ≈ 1d8+5 + 2d8 = mean ~14 per smitten attack.
  • A Ranger / Rogue with Sneak Attack. 1d8+5 + 5d6 sneak ≈ mean 27.
  • A blaster Wizard / Sorcerer landing 4d6 - 6d6 of damage per spell slot.

Aggregate party DPR sits roughly at 50-70 against Strahd's AC 16. The fight should resolve in 2-3 rounds — but variance, saves, and Strahd's regeneration / mist form complicate the simple division.

The renewal-theorem floor

For a single attacker with mean-per-swing μ against a target with HP H, the expected number of swings to drop the target is approximately:

E[N] ≈ H/μ + (Var/μ²)/2

Once H is comfortably bigger than your single-swing maximum (in Strahd's case, 144 ≫ 27 from a Greatsword crit), the variance correction is small and E[N] ≈ H/μ becomes a tight bound. For a Greatsword Fighter with GWM at mean per-swing 12 (without the Action-Surge round-1 boost), E[N] = 144/12 = 12 swings = 6 rounds (two attacks per round).

That's the per-attacker. With four PCs swinging simultaneously, expected rounds is approximately H / (sum of party means per round). At 60 DPR aggregate: 144/60 = 2.4 rounds. Plus variance.

Why crit-heavy compositions make this faster

The base 2d6+5 swing has mean 12. The crit-doubled version 4d6+5 has mean 19. With a 5% crit chance baseline and a Champion Fighter's improved critical (5-15%), the effective mean lifts to:

2d6+5
min 7 max 17 mean 12.00 12
P(finish at 14 HP) = 27.78% (5/18)
  • 7 2.78%
  • 8 5.56%
  • 9 8.33%
  • 10 11.11%
  • 11 13.89%
  • 12 16.67%
  • 13 13.89%
  • 14 11.11%
  • 15 8.33%
  • 16 5.56%
  • 17 2.78%
4d6+5
min 9 max 29 mean 19.00 19
P(finish at 14 HP) = 94.60% (613/648)
  • 9 0.08%
  • 10 0.31%
  • 11 0.77%
  • 12 1.54%
  • 13 2.70%
  • 14 4.32%
  • 15 6.17%
  • 16 8.02%
  • 17 9.65%
  • 18 10.80%
  • 19 11.27%
  • 20 10.80%
  • 21 9.65%
  • 22 8.02%
  • 23 6.17%
  • 24 4.32%
  • 25 2.70%
  • 26 1.54%
  • 27 0.77%
  • 28 0.31%
  • 29 0.08%

Champion's auto-crit-on-19+ effectively doubles the crit rate (from 5% to 10%). At 10%, an average swing's effective mean becomes 0.9·12 + 0.1·19 = 12.7 — about 6% higher mean DPR. Spread over 12 swings to drop Strahd, that's ~0.7 fewer expected swings = ~half a round saved.

Smite stacking changes the math more: on a crit, the smite dice double too. So a 2nd-level Smite that normally adds 2d8 (mean 9) doubles to 4d8 (mean 18) on a crit. Champion's doubled crit chance × Smite's doubled-on-crit dice = ~1.5x effective DPR on the round you smite.

The strikes-to-kill chain

The exact distribution of "rounds to kill Strahd" is what the strikes-to-kill chain computes. For a single attacker (the Greatsword Fighter with two attacks at 2d6+15 GWM = mean 22 per swing, plus the −5 hit penalty bringing effective mean to ~70% × 22 = 15.4 per swing):

2d6+15
min 17 max 27 mean 22.00 22
  • 17 2.78%
  • 18 5.56%
  • 19 8.33%
  • 20 11.11%
  • 21 13.89%
  • 22 16.67%
  • 23 13.89%
  • 24 11.11%
  • 25 8.33%
  • 26 5.56%
  • 27 2.78%

The strikes-to-kill panel (visible at the bottom of any strike page) shows the cumulative kill probability per attack-cycle for any HP target — including 144. For Strahd specifically, a 2d6+15 Fighter with two attacks per round, Action Surge on round 1 (four swings), and the rest a steady two-per-round, hits 50% kill probability around round 5 and 90% by round 8 — assuming Strahd doesn't regenerate, doesn't go to mist form, and the Fighter never misses. Real-world number is probably 4-6 rounds with a typical party.

The variance lesson

Strahd's 144 HP sits comfortably above any single PC's single-swing maximum. That puts us in the regime described in expected strikes to kill where mean wins linearly and variance becomes a small correction term. Practical implications:

  • High-variance builds matter less here. The Sorcadin's nova rounds and the Sneak-Attack Rogue's spike damage feel impressive, but over the course of the fight, steady-state DPR dominates. A reliable 60-DPR party kills Strahd faster than a bursty 70-DPR party that's hostage to crit fishing.
  • Action Surge shifts the curve. Front-loading damage in round 1 (when the Sorcerer hasn't run out of slots, when nobody's down yet) is worth more than the same damage spread across rounds. Burn the Surge first.
  • Don't chase the 1-round-finish fantasy. Even at 70 DPR, the variance around 144 HP makes a single-round kill statistically improbable without huge damage swings (a Sorlock burst, a stacked Crit Fish round). Plan for 3-4 rounds.

Try it yourself

↦ /kill/2d6+15/144 — single GWM swing vs Strahd ↦ /strike/2d6+15 — full GWM Greatsword distribution ↦ /strike/4d6+5 — crit version ↦ /vs/2d6+5/2d6+15 — base vs GWM, side-by-side

Adjacent reading: Expected strikes to kill — the renewal-theorem result this page applies to a specific boss. Great Weapon Master break-even — the +9 to-hit Fighter against Strahd's AC 16 sits above the GWM break-even, so GWM is on; the math from that pillar feeds straight into this one.