What it is. Structured replay. Prestige wipes your run for a permanent multiplier that makes the next climb faster; NG+ restarts the whole game harder with your build and gear carried over. Same content, re-authored decisions.

Player fantasy / why it’s fun. “I could tear through the part that broke me last time — and now the game finally fights back.” The reset stings for a second, then the acceleration is pure dopamine. Content you already own becomes a playground.

When to use / when NOT

Use it when:

• Your loop is short and repeatable — a run, a wave, a board. See process-core-loop.
• You have a scarcity of hand-authored content but a rich decision space that recombines.
• Mastery outpaces content: skilled players finish and want the game to escalate, not end.

Do NOT use it when:

• The first playthrough is a one-shot story whose surprises die on replay (a mystery, a twist) — a second lap is hollow. See genre-narrative-decisions, anchor-return-of-the-obra-dinn.
• You’d bolt it on to hide a thin core. Replay amplifies whatever the loop already is; a dull loop replayed is dull twice. See antipattern-content-desert.
• The metagame is already solved — one optimal path means every reset plays identically. See antipattern-solved-metagame.

Variants

Prestige answers “go again, faster.” NG+ answers “go again, harder.” Ascension answers “go again, on your terms.” Pick the question your loop earns.

Tuning levers

• What carries. The single most important knob. Carry too much → the next run is a victory lap (antipattern-fake-choice). Carry too little → the reset feels like punishment. Carry capability and knowledge, gate power.
• What resets. Reset enough that early decisions matter again. If you keep your endgame build, the opening 30 minutes are dead air. Slay the Spire keeps your card pool available but not assembled — you still draft.
• The new-run hook. Every reset must open a decision the last run couldn’t. New tier, new modifier, new enemy placement, a fresh seed. No new decision = antipattern-boring-optimal.
• The multiplier curve. Prestige gain must undercut the grind it replaces or nobody prestiges; overshoot and you trivialize the loop. Tune so a reset pays for itself in a fraction of the prior run’s time.
• Escalation slope. NG+ difficulty should re-threaten a mastered player without a wall. See system-difficulty-and-dda, antipattern-difficulty-cliff.
• Opt-in vs forced. Ascension is voluntary self-challenge; NG+ is often gated behind a first clear. Voluntary reads as respect; forced resets read as antipattern-grind-wall.
• Prestige floor. Guarantee each reset leaves the player strictly ahead. A reset that can net lose progress corrodes trust — see pattern-fairness-and-trust.

Twist seams

Two ways to break the “same content, again” tautology:

• NG+ but the world remembers your last run (twist: structure). Your previous playthrough writes state into the new one — the town you razed is ash, the boss you spared returns as an ally, doors mark where you died. Shades of anchor-shadow-of-mordor‘s Nemesis and anchor-hades‘s inter-run dialogue. Replay stops being a copy and becomes a sequel to your own choices. See world-level-as-story, pattern-escalation-and-payoff.
• Prestige but you keep knowledge, lose power (twist: constraint). The reset strips your stats but never what you learned — the map, the boss tells, the optimal route stay in your head, not the save file. The run gets mechanically harder yet subjectively easier because you got better. anchor-outer-wilds and anchor-dark-souls live here: the only thing that truly carries over is the player. See system-mastery-curve, pattern-mastery-and-flow.

How it wires to Hayao

• State partition. Split persistent state (meta-currency, unlocks, prestige tier) from run state (level, HP, board) so a reset clears one and keeps the other. Model the run as a discrete State your loop advances; the reset is a pure transition to a fresh initial state seeded by the persistent half. Compose with system-save-and-checkpoint.
• Determinism. New-run variety must come from your deterministic RNG, not wall-clock entropy — same seed + same carryover = same run, which is what lets you machine-prove an NG+ level is still winnable. See system-procgen-design and the sandboxes/procgen-lab.
• Difficulty as data. Express ascension/NG+ modifiers as a data layer applied over the base ruleset (enemy HP scalar, spawn table swap, added clauses) rather than a forked codebase — one game, many parameterizations. See system-difficulty-and-dda, system-loot-tables, system-spawn-directors.
• Prestige as an economy. The multiplier is a system-economy variable; tune its curve like any system-resource-loops sink/source and pressure-test with system-reward-schedules.
• Meta spine. This system is the replay engine; the persistent tree it feeds is system-meta-progression. Keep them distinct — meta-progression is what unlocks, prestige/NG+ is why you reset to earn it.

Fails when…

• Replay is identical — no new decision, no new threat, just a longer number. Dead loop. See antipattern-boring-optimal, antipattern-content-desert.
• The optimal reset path is solved and forced, so every prestige is the same rote sequence. See antipattern-solved-metagame, antipattern-fake-choice.
• Carryover is so strong the new run is a cakewalk — power creep with no counter-pressure. See antipattern-power-creep, antipattern-stat-inflation.
• The reset is a grind gate in disguise: the only way forward is to prestige, and prestiging is a slog. See antipattern-grind-wall, antipattern-pay-to-skip.
• NG+ spikes into a wall a mastered player still can’t clear. See antipattern-difficulty-cliff.

Verify

Prove it’s genuine longevity, not a treadmill: does each reset open a new decision and stay winnable? Assert on run/meta state via the loop’s probe; judge escalation against design/FUN.md#14-·-incremental.