pocket
Star
Fork
Watch
pocket docs vision.md
9.1 kB · 336 lines
Markdown
at main

Turn-Based Rhythm Roguelite — Working Combat Spec #

Pitch #

A focused, combat-first roguelite built around classic Final Fantasy-style party-vs-enemy battles, but with a rhythmic performance layer at the heart of every fight.

Less:

  • Guitar Hero
  • Note charts/lanes

More:

  • Turn-based RPG
  • build synergy / emergent gameplay
  • rhythmic combat

Feeling:

A speedrunner playing Slay the Spire crossed with Expedition 33, driven by incredible backing tracks, expressive party builds, and fast call-and-response combat.

Design Pillars #

Combat is built around "phrases" which guide the flow of combat seamlessly between planning and execution.

1. Rhythm Enhances #

Rhythm should feel additive, not busywork/reaction tax.

The difficulty is not in performing complex Guitar Hero-style note charts, rather it is matching your performance to your build, your party, enemies, and combat state.

A simple build should be playable with simple timing. A high-risk build can ask for more precision, faster reactions, off-beat execution, or longer phrase chains.

2. Strategic Between Phrases, Rhythmic Inside Phrases #

Preserve RPG strategy: party composition, jobs, passives, target priority, stagger, resources, defensive planning, and synergy.

But once a phrase begins, combat should move. The player should be reacting, guarding, countering, chaining, and recovering.

Avoid long, detached planning turns that break the musical flow.

3. Act and Response #

An enemy signals an action and the player responds by guarding, dodging, parrying, interrupting, countering, exploiting recovery, or setting up a later payoff.

Mastery comes from learning enemy rhythm, reading body language, feeling the music, and using the party build to answer threats.

4. Builds Change How You Perform #

Jobs, characters, relics, and passives alter the player’s relationship to rhythm.

One build may care about downbeats. Another may care about off-beats. Another may reward perfect guards. Another may automate support actions. Another may extend combo windows. Another may create delayed spell detonations.

The question is not only:

What does my build do?

But also:

How does my build want to be played?

5. Pressure Without Panic #

The player should feel encouraged to act quickly and stay in flow.

The game should not require chess-clock stress at all times, but combat should have enough implicit or explicit pressure that overthinking feels suboptimal.

The ideal feeling is fast, confident decision-making under rhythmic momentum.

Target Feel #

  • classic JRPG battle presentation
  • roguelite build escalation
  • rhythmic combat mastery
  • Expedition 33-style timing and enemy reading
  • Slay the Spire-style run decisions
  • speedrunner-like flow
  • party synergy as performance
  • fights that feel like songs without being “pick a song” rhythm stages

Anti-feel #

  • Guitar Hero with RPG damage numbers
  • a pure action game
  • a slow tactics game
  • a menu-heavy JRPG
  • a note highway memorization game
  • a game where every single action demands precise timing
  • a game where rhythm mastery invalidates build strategy
  • a game where build strategy invalidates execution

Core Combat Loop #

Each battle flows through repeated combat phrases.

At a high level:

  1. The player chooses or confirms party actions.
  2. Actions enter a rhythmic performance window.
  3. Enemies perform attack phrases.
  4. The player defends, interrupts, counters, or exploits openings.
  5. Damage, stagger, tempo, buffs, debuffs, and passives resolve.
  6. The fight continues without fully losing musical momentum.

Combat feels like alternating pockets of micro-planning and rhythmic execution.

Measures and Beats #

Internally, combat is organized around beats and measures.

Most early combat should probably use 4/4 timing for readability.

Example internal structure:

Beat:    1       2       3       4
Enemy:   windup  strike  recover idle
Player:  choose  guard   counter chain

This does not need to be displayed as a literal grid. The player should feel it through animation, sound, VFX, UI pulse, and hit-stop.

The beat system exists to make actions legible, authorable, and tunable.

Composure #

Enemies have Composure (stagger) meter.

This is the bridge between rhythm execution and RPG strategy.

The player reduces enemy Composure by:

  • perfect attacks
  • parries
  • interrupts
  • exploiting recovery
  • hitting weaknesses
  • using job-specific tools
  • answering enemy phrases correctly

When Composure breaks:

  • enemy loses or delays its next phrase
  • enemy takes bonus damage
  • enemy exposes a vulnerability
  • party gains Tempo
  • certain job passives trigger
  • combo routes open

This allows enemies to have rhythmic weaknesses.

Tempo #

Tempo is the shared party flow resource (mana).

Tempo represents whether the party is in control of the performance.

The party gains Tempo from:

  • perfect actions
  • exploiting recovery windows
  • clean combo chains
  • job synergies
  • defeating enemies on beat
  • maintaining streaks

The party loses Tempo from:

  • misses
  • failed actions
  • unguarded hits
  • interrupted casts
  • broken combo chains
  • panic actions
  • certain enemy attacks

Tempo might power:

  • job abilities
  • emergency defenses
  • extra actions
  • ultimate attacks
  • party assists
  • rhythm window expansion
  • gambit triggers
  • relic effects

Tempo should create emotional battle arcs. At high Tempo, the party feels unstoppable. At low Tempo, the player can still recover, but must play more carefully.

Action Economy #

The game needs enough pressure to avoid detached over-planning.

Need to prototype different ideas here and see what works.

Characters #

A character has:

  1. A Job
  2. Passive(s)

A Job has:

  1. Stance(s)

A Stance has:

  • An Ability
  • Gambit(s)

Jobs #

Jobs provide abilities and gambits and can be upgraded. Jobs are the power milestones for characters.

Stances #

Stances are combat modes granting an ability and gambit(s).

  • stances can be "swapped" in combat to cycle from a set of abilities to another
  • swapping stances is an action with a cast time (usually short)
  • swapping stances interrupts actions

Passives #

Innate, permanent effects.

Abilities #

Abilities are actions that may be manually activated by the player.

  • has a cast time (a "windup" period prior to firing) and effect(s) (the outcome of the ability, e.g. deals 5 damage)
  • if cast time is 0, the effect fires on the beat that the player activates it
  • can either be tapped (single press) to cast immediately or held (hold button down) to begin casting and released to fire at a player-determined time
  • can be interrupted (by enemies and self via stance-swapping)
  • can be input buffered

Gambits #

Gambits are actions that cast automatically if no other action is casting

  • may have pre-requisites before activating (e.g. health < 50%)
  • can be interrupted

Party Structure #

A party contains 1-3 characters.

This keeps characters readable and simple while allowing jobs, relics, and upgrades to create build variety.

Roguelite Structure #

The overworld uses node-based movement similar to Slay the Spire.

The combat is the primary focus. The map exists to create run structure, risk/reward decisions, and build variation.

Boss design #

Bosses should feel composed, but not like selected songs.

A boss is a combat musician with a motif.

The fight should teach a phrase, mutate it, then test mastery.

The player feels progress because they recognize the pattern, then adapt to variations.

The best boss fights should feel like learning a duet.

Music Direction #

The backing music is integral, but fights are not “pick a song.”

Music should support readable timing, emotional momentum, and combat identity.

Important principles:

  • strong beat readability
  • clear attack accents
  • adaptive layering
  • enemy phrases supported by instrumentation
  • Tempo state reflected in music intensity
  • boss phases reflected in arrangement
  • perfect hits add musical punctuation
  • misses create subtle disruption
  • high Tempo creates richer mix layers

Co-op Potential #

Potential future system.

Co-op could work if each player controls one or more party members.

The appeal:

  • each player performs their character’s rhythm phrases
  • party synergy becomes actual coordination
  • one player guards while another casts
  • one player sets up stagger while another bursts
  • boss fights become ensemble performances

Accessibility #

Because rhythm and timing are core, accessibility needs to be considered early.

Ideas:

  • widen timing windows
  • reduce planning pressure
  • visual beat assist
  • audio beat assist
  • input latency calibration
  • auto-guard assist
  • simplified defensive mode
  • reduce flash/VFX intensity
  • practice mode for enemy phrases
  • separate rhythm difficulty from strategic difficulty

The game should remain interesting when timing is easier, because build strategy, enemy reading, and resource management still matter.