Guide

Game install super and install mode systems explained

Harbor Brawl's rushdown anchor shipped with a three-bar “Overdrive” super: 40 frames of startup, a full-screen explosion, then back to the normal movelist. Telemetry showed players saved it for round-end confirms — only 11% of neutral wins involved Overdrive, and the character's win rate lagged the roster by six percentage points. The problem was not damage; it was shape. A burst super spends meter once and ends. An install super (or install mode) spends meter to enter a sustained powered state: altered normals, movement buffs, new specials, and a visible timer that drains while you press the advantage. After refactoring Overdrive into a 12-second install with +12% damage, faster dash, and install-only rekka routes, install activation rose from 19% to 47% of winning rounds and neutral conversion improved without raising single-hit burst damage.

Install mode is the combat-design pattern where a character temporarily becomes a different kit — not a cinematic one-shot, not a permanent stance toggle, but a timed transformation gated by super meter. This guide covers install taxonomy, drain models, movelist morphing rules, punish windows, interaction with burst escapes, the Harbor Brawl refactor, a technique decision table versus burst supers and stance systems, pitfalls, and a production checklist for teams building readable power spikes that reward sustained offense rather than single-button wins.

What install mode is (and what it is not)

Install mode is a combat state entered by spending super meter (or a dedicated install gauge) that lasts until a timer expires, the player is knocked down under specific rules, or a secondary spend condition triggers. While active, the character receives buffs and often a partial or full movelist swap. The player's job shifts from “land one super” to “maximize value before the aura ends.”

Not the same as a burst super

Burst supers are single transactions: input, animation, damage, state ends. Install supers are subscriptions: you pay upfront, then fight inside the powered window. Burst supers pair with cinematic freeze presentation; install modes pair with persistent UI (timer bar, aura VFX, altered idle animation).

Not the same as stance switch

Stance switches are usually free or low-cost and reversible at will. Install mode is meter-gated, time-limited, and often cannot be cancelled early without penalty. Stance design optimizes for long-term character identity; install design optimizes for comeback tempo and round-swing moments.

Not the same as rage or desperation mode

Desperation triggers at low HP and may persist until round end. Install is player-initiated and costs resources the attacker chose to bank. Mixing the two without clear UI creates confusion about whether buffs are earned or automatic.

Install taxonomy: how powered states end

The drain model defines pacing. Pick one primary clock and document exceptions in the movelist CSV.

Fixed timer. Install lasts 8–15 seconds regardless of action. Simple to read; encourages immediate offense. Guilty Gear-style installs often use this with visible countdown.
Hit-extending timer. Each successful hit adds 0.3–0.8 seconds up to a cap. Rewards conversion loops; risks infinite pressure if cap is too high.
Hit-draining timer. Getting hit removes 2–4 seconds or ends install on hard knockdown. Creates clear counterplay: survive, punish, wait it out.
Meter drain per second. Install consumes fractional super bars while active; ending early refunds none. Ties install to the same economy as EX moves.
Action spend. Each install-only special costs a slice of remaining install time. Prevents spam of install-exclusive tools.

Hybrid models work when documented: Harbor Brawl uses fixed 12-second base timer, −3 seconds on counter-hit during install, +0.5 seconds per normal hit (cap +4 seconds total extension).

Movelist morphing and buff packages

Install must change how the character plays, not only numbers on the same buttons. Common morph patterns:

Buff-only install

Same inputs, better data: +10–20% damage, faster movement, improved frame advantage on key normals. Lowest implementation cost; can feel invisible if VFX is weak.

Route swap install

Normals chain into install-only follow-ups; specials gain new properties (wall carry, launcher, projectile nullify). Requires duplicate move entries in the moveset table with requires_install flag.

Replacement install

Entire button map changes (e.g. punch becomes rekka, kick becomes teleport). Highest skill ceiling; demands training mode overlays showing install mappings.

Typical buff budget

Damage multiplier: 1.08–1.20 (avoid stacking with install-only combo scaling bonuses).
Movement: 10–25% dash speed or air mobility.
Defense: optional super armor on install-only moves only — not on every normal.
Meter: install may disable further super spend or allow one install-exclusive finisher that ends the state.

Startup vulnerability and punish design

Install activation is the risk payment. If startup is too fast, players install mid-combo with no counterplay. If too slow, the tool never fires at neutral. Production targets:

Startup: 25–45 frames with standing hurtbox; throwable and launchable on whiff.
Invincibility: avoid full invuln unless install is a reversal-only tool; 0–8 frames upper-body invuln is typical.
Block advantage on install trigger: if install cancels from blockstring, make it minus enough that mash or reversal wins unless conditioned.
Recovery on whiff: 20+ frames launchable; opponents must be rewarded for baiting install startup.

During install, the attacker should not be invulnerable. Burst, throw, and hard knockdown remain valid outs. If install persists through soft knockdown, oki becomes oppressive — most games end install on HKD or force stand the defender with install timer still running (designer choice, but pick one).

UI, audio and training-mode contracts

Players must read install state at a glance:

Timer bar adjacent to health (not hidden in super meter).
Persistent aura tint on character and hit sparks.
Distinct install idle and walk cycle.
Audio sting on enter and exit; pitch shift on last 3 seconds.
Training overlay: install remaining time, extension events, morph input map.

Rollback netcode must sync install state as a first-class flag — desynced install timers produce “ghost install” reports in online play. Serialize install start frame, remaining duration, and extension cap in the battle snapshot.

Harbor Brawl refactor: burst super to install mode

Pre-refactor Overdrive (burst):

3-bar cost, 420 damage, 40f startup, used mainly after HKD confirm.
Install activation rate in winning rounds: 11%.
Character win rate: 44% (roster average 50%).

Post-refactor Overdrive (install):

2-bar entry, 12s base timer, +12% damage, +18% dash speed, rekka 5P route.
Install activation in winning rounds: 47%.
Character win rate: 49%.
Round swing after successful install entry: +28% win probability (internal sim).
Burst finisher removed; install-exclusive super ends state and costs remaining timer.

Critical tuning: lowering entry to 2 bars without shortening install startup (kept at 32f) prevented install-from-plus-frames abuse while making neutral install a readable gamble.

Technique decision table: install vs alternatives

Power spike type	Duration	Best for	Weak when
Burst super (one-shot)	Single hit	Confirm damage, cinematic payoff	Sustained neutral pressure needed
Install mode	Timed window	Conversion loops, comeback tempo	Defender can stall or burst out
Stance switch	Until toggled	Long-term archetype duality	Round-swing moments need meter cost
Transform (e.g. boss phase)	Often permanent per round	Character fantasy, raid mechanics	Competitive balance across roster
Rage / desperation	Low-HP trigger	Automatic comeback assist	Skill expression on resource spend
EX move tier	Per input	Incremental meter sinks	Defining character power identity

Common pitfalls

Install with no movelist change. +15% damage alone rarely justifies 2–3 bars; players prefer guaranteed burst.
Zero startup vulnerability. Neutral becomes install roulette; defenders cannot check activation.
Timer hidden in UI. Opponents cannot plan stall or burst timing; feels opaque online.
Install persists through hard knockdown. Oki loops exceed juggle budgets and generate infinite pressure reports.
Uncapped hit extension. Timer never expires if player keeps landing jabs; add hard cap (+3–5s).
Stacking with other buffs. Install + rage + counter-hit damage multiplier produces one-touch round wins.
No install exit telegraph. Sudden return to weak normals mid-combo confuses; fade aura 1s before end.
Identical install across roster. Removes character identity; limit to 1–3 install archetypes.

Production checklist

Define install drain model (timer, hit-drain, meter-drain) in design doc.
Tag all install-only moves with requires_install in moveset DB.
Budget 25–45f vulnerable startup on install activation.
Cap hit-based timer extension (+3–5s typical).
Document HKD/soft KD behavior: install ends or persists?
Expose timer bar and aura VFX readable at 1080p and mobile scale.
Sync install state in rollback snapshots and replay files.
Training mode: install input map overlay and extension event log.
Telemetry: entry rate, average damage during install, timeout vs KO during install.
Limit install characters to clear fantasy (rushdown, zoner install trap, etc.).
Playtest stall strategies: backdash, burst, projectile wall at timer end.
Balance entry cost (1–3 bars) against expected install damage ceiling.

Key takeaways

Install mode trades burst damage for a sustained powered window — the design goal is conversion tempo, not one-button round theft.
Drain model (fixed timer, hit extension, hit penalty) defines counterplay as clearly as frame data on the install trigger.
Movelist morphing or route swaps make install feel transformative; numeric buffs alone underperform in telemetry.
Vulnerable startup and visible timer UI are mandatory for fair neutral and online readability.
Harbor Brawl's burst-to-install refactor raised activation from 11% to 47% of winning rounds without increasing single-hit burst damage.