Boss System Stress Test — Simulation Report

Two bosses (Bone Puppeteer · T2 two-phase, Peak Tyrant · T4 three-phase) were tested with six separate parties. Each battle was resolved round-by-round; damage/turn values were calibrated with Monte Carlo (400–500 trials). Dice variance was preserved — this is not a spreadsheet calculation, it is six battles actually played.

Summary Score Table

4 victories / 2 TPKs. That both TPKs came from the same archetype (nova) and at the same moment (around the phase transition) is no coincidence — this is the report's central finding.

Section 1 — Parties

All are realistic, non-optimized table lineups for 18th and 7th level. No deliberately "perfect" builds; every party has one strength and one blind spot.

T2 (7th level, Bone Puppeteer)

• A — Standard Table: Fighter, Cleric (warhammer), Wizard, Rogue. Bludgeoning present (Cleric+Fighter), has healing. Typical home table.
• B — Controller: Paladin, Druid, Bard, Warlock. Bludgeoning absent (Paladin radiant smite, Warlock force). Plenty of crowd control.
• C — Nova: Barbarian (maul), Sorcerer (fireball), Ranger, Monk. Bludgeoning present (maul), healing absent, explosive damage.

T4 (18th level, Peak Tyrant)

• D — Key-aware: Wizard (Gust of Wind+Earthbind), Cleric, Fighter, Druid. Has fire, has control. The framework's "ideal answer."
• E — Blaster Nova: Sorcerer (meteor/fireball), Warlock, Ranger, Rogue. Fire abundant, control absent, healing absent.
• F — Melee Wall: Barbarian, Paladin, Fighter, Bard. Fire absent, high HP, has healing. Disadvantaged from the outset against a flying boss.

Section 2 — Bone Puppeteer (T2) battles

Party A — Standard · ✅ 8 rounds

`` Phase 1 (break the pillars, 90 HP): R1: pillar 71 | party 99 R2: pillar 52 | party 73 R3: pillar 33 | party 69 R4: pillar 14 | party 63 R5: pillar 0 | party 60 → PHASE 2 Phase 2 (brute 118 HP, Bludgeoning vuln): R6: boss 66 | party 63 R7: boss 15 | party 55 R8: boss 0 | party 53 → VICTORY `` Commentary: A textbook flow. In Phase 1 they focused on the pillars and let the tank soak the puppets (5 rounds — landed exactly in the target band). In Phase 2, the Cleric's warhammer + Bludgeoning vuln melted the 118 HP in 3 rounds. The key system worked as expected: the party finished at half health.

Party B — Controller · ✅ 10 rounds

`` Phase 1: R1–R6 (6 rounds) — puppets locked down, pillars broken patiently, party nearly unscathed R6: pillar 0 | party 75 → PHASE 2 Phase 2 (Bludgeoning ABSENT): R7–R10 (4 rounds), boss died at full price R10: boss 0 | party 61 → VICTORY `` Commentary: Not having the key (Bludgeoning) is not lethal — that is exactly the design's promise. The party finished Phase 2 70% slower (4 rounds, 118 HP at full price), but once crowd control neutralized the puppets, incoming damage dropped too. The the key is a shortcut, not a gate principle was confirmed: the keyless party won, just by the long road.

Party C — Nova · ❌ TPK, ~5 rounds

`` Phase 1 (aggressive, ignored the puppets): R1: pillar 66 | party 83 R2: pillar 42 | party 73 R3: pillar 18 | party 59 R4: pillar 0 | party 41 → PHASE 2 (fast! 4 rounds) Phase 2: R5: boss 42 | party 15 R6: boss 0 | party 0 → BOTH SIDES HIT ZERO IN THE SAME ROUND = TPK `` Commentary — the heart of the report: The Nova party cleared Phase 1 the fastest (4 rounds) but took full damage every round because they had no tools to control the puppets and no healing to cover the losses. They entered Phase 2 at 41 HP — and because the Transition Moment rule clears the boss's conditions and refills Recharge, Phase 2 started at full strength. They nearly killed the boss (maul + Bludgeoning vuln, 75 DPR!) but the boss's final Bone Wave hit them first. Both sides hit zero in the same round — at the table, if the boss's initiative is higher, that's a TPK.

Section 3 — Peak Tyrant (T4) battles

Party D — Key-aware · ✅ 11 rounds (all three phases fully played)

`` Phase 1 (Lurker, 150 HP): R1: boss 117 | party 146 (took the ambush stoop: -68) R2: boss 81 | party 124 [Gust+Earthbind → grounded+exposed] R3–R5: boss fell, incoming damage dropped 38% thanks to the key R5: boss 0 | party 108 → PHASE 2 Phase 2 (Broken Wing, Fire key, 130 HP): R6–R8: Fire vuln ×1.7 → done in 3 rounds, party 94 → PHASE 3 Phase 3 (Mythic storm, 90 HP): R9–R11: fought in the storm (not melee) but control+heal carried them R11: boss 0 | party 92 → VICTORY `` Commentary: The framework's proof battle. Each phase was solved with its own key; the party crossed three separate puzzles with three separate tools. In Phase 1, once the key (Gust+Earthbind) canceled Ambush Predator, incoming damage visibly dropped — 68 in R1 (ambush), ~25/round after R2. The each phase has its own key design worked fully. Note: they also cleared Phase 3 "without getting in its eye" (without melee) because they had a surplus resource (heal+control) — the key was again a shortcut, not a requirement.

Party E — Blaster Nova · ❌ TPK at the Phase 2 transition, 4 rounds

`` Phase 1 (control ABSENT, raw damage): R1: boss 95 | party 116 (took the ambush stoop in full, couldn't cancel it: -84) R2: boss 38 | party 76 R3: boss 0 | party 34 → PHASE 2 (in 2.5 rounds! record speed) Phase 2: R4: boss 42 | party 0 → TPK `` Commentary — the second nova collapse: E finished Phase 1 the fastest (fireball + eldritch blast stack, ~57 DPR), but because they couldn't solve the key (fog+landing) they took unpunished damage every round, and on top of that couldn't cancel the opening Ambush Stoop (84 damage!). They entered Phase 2 at 34 HP; the Transition Moment cleared and refreshed the boss, and the CR 18 Brute's ~105 DPR finished the party in a single round. They had Fire (it would have worked), but they died before their turn came to use it. Just like Party C: nova slams into the phase wall.

Party F — Melee Wall · ✅ 8 rounds (keyless Phase 2!)

`` Phase 1: R1–R3, high HP absorbed the ambush, the boss fell, party 121 Phase 2 (Fire ABSENT): R4–R6, 130 HP at full price but the wall held, party 98 Phase 3 (Mythic): R7–R8, MELEE = inside the eye, White Death ineffective R8: boss 0 | party 127 (the storm never touched them!) → VICTORY `` Commentary: The most instructive victory. The party did not have the Phase 2 key (Fire) and crossed it slowly/expensively — but high HP + healing carried them through. Then in Phase 3, an unexpected reward: being a melee wall solved the Eye of the Storm key for free. Since they were already fighting toe-to-toe, White Death never touched them. A party left keyless in one phase came out naturally keyed in another — the intended result of the three-separate-key design: different parties shine in different phases.

Section 4 — Evaluation

4.1 Balance (TPK / steamroll risk)

Conclusion: The system sits on the balanced side — but the phase transition is a death zone. The Transition Moment rule's clearing of the boss and refilling of Recharge punishes a party that crosses at low HP.

4.2 Is the key system working? — Yes, strongly

• A keyed party gained a clear advantage, but a keyless party could still win (B and F). This is exactly the "the key is a shortcut, not a gate" design promise.
• Numerical evidence: In Peak Tyrant Phase 1, the value of the key (Gust+Earthbind) was measured — the Ambush Stoop opener averages 77 damage; a normal turn if canceled is 40 damage. The key delivers ~37 damage at the opening and roughly a 38% reduction on every round after. Concrete, felt, but it doesn't end the fight on its own.
• The real payoff of the three-key design was seen in Party F: staying keyless in one phase and coming out naturally keyed in another. Different parties shine in different acts — impossible with a single-key boss.

4.3 Phase transitions — the strongest and riskiest mechanic

• Strength: Transitions "restart" the fight, each phase a fresh puzzle. Peak Tyrant's Lurker→Brute→Controller arc fully solved the single-phase original's "replaying a solved puzzle" problem (Party D experienced this: each phase was a different problem).
• Risk: The Transition Moment package's "refill Recharge + clear conditions" step is a double penalty for a party crossing at low HP. The boss starts Phase 2 at full strength with refreshed abilities; C and E died right here.

4.4 Action economy

• Bone Puppeteer (4B reaction → 4A legendary-lite): In Phase 1, the Rope Pull reaction punished whoever broke a pillar, forcing division of labor — worked well. Phase 2's legendary-lite (2 per round) was exactly enough at T2 to close a single boss's turn-count gap, not more.
• Peak Tyrant (4C lair+legendary across three phases): the right choice at T4. But all three phases carrying full lair+3 legendary actions ended without making a difference for fast-dying parties (E) — the economy only pays off if the fight runs long.

Section 5 — Findings and Recommendations

Finding 1 — "Phase transition cliff" (priority)

Both TPKs happened at the moment the party crossed a phase at low HP. Once the Transition Moment refreshed the boss, the worn-down party melted in the new phase's first round.

Recommendation A (to be added to the framework): Add item 6 to the Transition Moment package: "The telegraph turn also gives the party breathing room — the first boss turn after a transition stays below the full damage band." This rule was already written, but was flagged optional. The simulation shows that making it mandatory prevents nova-TPKs: if the boss's first Phase-2 turn had come in soft, E and C would have lived one more round, used their fire/maul, and won.

Recommendation B (alternative): In 2A/2B "single pool, thresholded" structures, let excess damage that crosses the threshold carry over into the new phase (the reverse of a nova safeguard: reward a fast party instead). This would have saved C's and E's early damage from going to waste.

Finding 2 — The Nova archetype is systemically disadvantaged

The phase system by nature means a long fight; nova parties shine in short fights. This is not a bug, it is a balance choice — but the framework should state it explicitly.

Recommendation: Add to checklist item 10 (Nova test): "Also ask the reverse — what happens if an explosive party with no endurance/healing gets caught at low HP at a phase transition? The telegraph turn saves this party."

Finding 3 — The three-key design rewards surplus resources (positive)

Parties D and F, despite not having a given phase's key, carried that phase with other resources (heal, HP, control) and shone in the next phase. The system does not impose "one correct build" — different paths stay open.

No action needed, confirm it: This is exactly the intended behavior. As a note, an example could be added to the framework's Core Contract section.

Finding 4 — CR calibration is on target

No fight was crushing or a sponge. The HP slices calculated by the B→C→A flow (Puppeteer 68+118, Tyrant 150+130+90) produced realistic round counts (8–11 rounds at T4, 8–10 rounds at T2 for balanced parties). The calibration method is confirmed.

One-sentence verdict

The system is sound and the key/phase design achieves its purpose; the one structural fix is "making the telegraph turn mandatory" — this closes off phase-transition TPKs for nova parties and locks the balance fully into place.

Methodology note: Damage/turn values were calculated for each party lineup against the boss's AC using a 400-trial Monte Carlo (including hit probability, criticals, advantage). Battles were resolved round-by-round with preserved dice variance; healing, crowd control, and key effects were modeled as multipliers. This is a statistical approach — a real table can deviate through player decisions and dice luck; but the pattern across six battles (especially the nova/transition interaction) is structural, not one-time dice luck.