Guide

Game stun and dizzy systems explained

Harbor Brawl's ranked telemetry after the blockstring patch showed a stubborn pattern: platinum matches averaged 94 seconds per round, and 41% ended by timeout rather than KO. Rushdown characters chipped health with blockstrings, but defenders who held block could stall indefinitely — there was no second loss condition beyond health. Designers tried raising chip damage; turtles simply walked away more. The fix was a stun gauge that filled on blocked and hit attacks, decayed when pressure stopped, and triggered a short dizzy state at 100% with a mash-out escape.

Average round length fell to 68 seconds; stun KOs rose from 0% to 14% of decisive rounds without increasing per-hit damage. This guide covers stun gauge math, dizzy state machines, mash-out and decay tuning, interactions with throws and counter-hits, the Harbor Brawl stun refactor, a technique decision table versus health-only rounds, pitfalls, and a production checklist.

What stun and dizzy systems are

A stun gauge (also called stun meter, dizzy meter, or guard crush meter in some titles) tracks how much punishment a fighter has absorbed in a round. When the gauge reaches its maximum, the defender enters a dizzy state — a brief vulnerability window where they cannot block and may take extra damage or guaranteed follow-ups.

Stun systems add a parallel win vector to health. They reward sustained pressure without requiring infinite combos or inflated chip damage. They also create visible tension: both players watch the stun bar climb during long blockstrings, which communicates “this round is ending soon” even when health is still high.

Stun is distinct from hitstun and blockstun (per-move freeze frames), from poise and posture (interrupt resistance in action games), and from guard-break meters (which break block rather than dizzy). Stun accumulates across multiple exchanges; hitstun resets every time a move connects.

Stun gauge accumulation

Every damaging interaction adds stun value to the defender's gauge. Common inputs to the stun function:

  • Hitstun contribution — base stun per move, often tied to the move's hitstun frames or a separate stun table.
  • Blockstun contribution — blocked attacks add a fraction of hit stun (typically 25–50%) so turtling still fills the gauge.
  • Counter-hit bonus — counter hits may add extra stun on top of damage, linking stun pressure to frame wins.
  • Throw stun — throws add a lump sum; command grabs may add more than regular throws.

Designers usually apply combo scaling to stun as well as damage: each successive hit in a combo contributes less stun than the first. Without stun proration, one long juggle could dizzy from zero; with proration, dizzy requires varied pressure across the round.

A typical stun formula per hit:

stun_added = base_stun[move] * stun_proration[hit_index] * counter_mult * block_mult

Where block_mult is 0.3–0.5 for blocked hits and 1.0 for clean hits. Counter mult is 1.0–1.25 depending on whether counter bonuses extend to stun.

Stun decay and recovery

If stun only climbed and never fell, the first long blockstring would dizzy every round. Stun decay drains the gauge when the defender is not taking stun damage — usually after a few seconds of neutral or when neither player is in blockstun/hitstun.

Decay tuning controls round pacing:

  • Fast decay rewards burst pressure; defenders recover between strings. Good for fast-paced anime fighters.
  • Slow decay makes stun sticky; one bad blockstring haunts the whole round. Good for methodical footsies games.
  • Decay pause during blockstun — gauge does not decay while defender is actively blocking, so chip-and-stun pressure stays threatening.
  • Decay acceleration at low health — optional catch-up mechanic so stun KOs do not dominate early while still closing stalled rounds.

Harbor Brawl uses 42 stun points per second decay after 1.2 seconds of neutral, with decay paused during blockstun. A full gauge is 100 points; average blockstring adds 8–12 before proration.

Dizzy state and mash-out

When stun reaches maximum, the defender enters dizzy (also called daze, stun lock, or guard crush in different franchises). Core properties:

  • Cannot block — all incoming attacks connect as if defender were standing idle.
  • Fixed duration or mash-out — dizzy lasts N frames unless the defender mashes buttons to reduce duration.
  • Visual telegraph — stars, swaying animation, or color shift so both players know the state started.
  • Reset on wake — gauge drops to 0% (or a high watermark like 80%) after dizzy ends; cannot immediately re-dizzy without new pressure.

Mash-out design

Mash-out lets defenders reduce dizzy duration by rapidly pressing buttons. Good mash-out design:

  • Caps minimum dizzy time (e.g. 30 frames) so attackers always get one confirm.
  • Uses diminishing returns so autofire macros do not trivially escape.
  • Displays a subtle mash meter so defenders know escape progress.
  • Does not require a specific button sequence — any face button counts.

Attackers get a guaranteed opening: a throw, a jump-in, or a command grab that ignores mash. Dizzy is a reward for pressure, not a free infinite.

Stun interactions with other systems

Stun does not exist in isolation. Document these cross-rules in your combat spec:

  • Throws during dizzy — usually guaranteed unless defender has a throw-tech window (rare during dizzy; most games disable tech).
  • Stun during juggles — airborne hits may add reduced stun or none, so juggle loops do not dizzy from one starter.
  • Stun on counter-hit crumple — crumple states may add bonus stun lump sum, linking counter routes to dizzy threat.
  • Stun immunity after dizzy — brief immunity window prevents infinite dizzy loops from the same string.
  • Super armor and stun — armored moves may still add stun to the defender even when damage is absorbed.
  • Round-end stun carry — most games reset stun each round; some tag fighters carry a fraction between rounds for drama.

Harbor Brawl stun refactor

Before stun, Harbor Brawl rounds had one dominant strategy: walk back, block, and wait for timer or a single opening. Rushdown characters felt unrewarding because their strength — sustained pressure — did not close rounds.

The refactor added:

  • 100-point stun gauge with UI bar under each health bar.
  • Per-move stun table separate from damage (fast lights add less stun per hit than slow heavies).
  • Stun proration matching damage proration curve from hit 3 onward.
  • 42 points/sec decay after 1.2s neutral; decay paused in blockstun.
  • 1.8-second dizzy at 100% with mash-out floor of 0.5 seconds.
  • Post-dizzy immunity for 4 seconds of match time.
  • Training mode stun overlay showing per-hit stun contribution.

Outcomes: timeout rate fell from 41% to 19%; average combo length unchanged; players reported higher tension during blockstrings without feeling “combo killed” by raw damage spikes.

Technique decision table

Goal Prefer stun/dizzy system Prefer alternative
Close long defensive rounds without damage inflation Stun gauge with blockstun contribution Higher chip damage only (often feels grindy)
Reward blockstring conditioning Stun decay paused during blockstun Guard-break meter (breaks block, not dizzy)
High-tension comeback swings Dizzy with mash-out and guaranteed opener Health-only with super meter comeback
Casual accessibility Visible stun bar + training overlay Disable stun in casual queues
Prevent one-combo dizzy Stun proration from hit 2+ Flat stun cap per combo segment
Anti-turtle without grabs Blockstun fills stun at 40%+ rate Command grabs only (see grab guide)

Common pitfalls

  • No stun proration. One juggle or infinite blockstring dizzies from zero; matches feel arbitrary.
  • Stun decay too fast. Gauge never fills; the system is cosmetic.
  • Stun decay too slow. First touch dooms the round; defenders feel helpless.
  • Invisible stun values. Players cannot learn which moves threaten dizzy without a published stun table.
  • Dizzy with no minimum duration. Mash-out at frame 1 makes stun KOs impossible to confirm.
  • No post-dizzy immunity. Re-dizzy loops from the same string breed frustration.
  • Blockstun adds zero stun. Turtles ignore the system entirely; only hit-confirms matter.
  • Dizzy during online rollback. Mash-out counters must be deterministic from input count, not frame-perfect client timing.

Production checklist

  • Define max stun gauge value and per-move stun table separate from damage.
  • Apply stun proration curve aligned with combo damage scaling.
  • Set blockstun stun multiplier (typically 0.3–0.5 of hit stun).
  • Author decay rate, neutral delay before decay, and blockstun pause rules.
  • Specify dizzy duration, mash-out floor, and mash diminishing returns.
  • Add post-dizzy immunity window and gauge reset watermark.
  • Document throw stun, counter-hit stun bonus, and juggle stun reduction.
  • Expose stun bar in HUD with color shift above 70% threshold.
  • Build training mode stun overlay with per-hit contribution log.
  • Test timeout rounds: does stun close stalls within target round length?
  • Validate mash-out under rollback netcode simulation.
  • Publish stun table alongside frame data for tournament players.

Key takeaways

  • Stun gauges add a second win vector that rewards sustained pressure without inflating per-hit damage.
  • Blockstun contribution is essential — turtles must fill the gauge while blocking or the system fails.
  • Stun proration and decay tune whether dizzy comes from one combo or a full round of conditioning.
  • Dizzy needs a mash-out floor and post-dizzy immunity so confirms are fair but not infinite.
  • Harbor Brawl cut timeout rates by 22 points by making blockstrings threaten dizzy, not just chip.

Related reading