pn-game-logic

Game logic skill

When to use

• Implementing or refactoring game loop architecture
• Managing game states (menu, play, pause, gameover)
• Handling input (keyboard, mouse, gamepad)
• Implementing collision detection or spatial partitioning
• Adding scoring, progression, or save/load
• Structuring game logic for testability

Game loop

1. Fixed timestep: Use a fixed delta (e.g. 1/60s) for physics and deterministic logic. Accumulate elapsed time and step in fixed increments to avoid spiral of death.
2. Accumulator pattern: accumulator += delta; while (accumulator >= FIXED_STEP) { update(); accumulator -= FIXED_STEP; }. Cap accumulator to avoid large catch-up steps.
3. Render interpolation: For smooth visuals, interpolate between last and current physics state using accumulator / FIXED_STEP when rendering.
4. Single loop: Use one render loop with requestAnimationFrame. Separate update(delta) (logic) from render() (draw).

State machines

• Game states: Model high-level states (menu, play, pause, gameover, settings). Use explicit state enum or state machine (e.g. currentState, enter(), exit(), update()).
• Transitions: Define valid transitions; guard against invalid state changes. Use events or callbacks for state entry/exit (e.g. play music on play, pause physics on pause).
• Sub-states: For complex flows (e.g. play → cutscene → play), use nested or hierarchical state machines when needed.

Input handling

• Action mapping: Map raw keys/buttons to logical actions (e.g. "jump", "move", "shoot"). Decouple input from game logic for rebinding and multiple devices.
• Input buffering: For responsive controls, buffer inputs for a short window (e.g. jump buffer so late jump still registers). Clear buffer on consume.
• Gamepad: Normalize axes and buttons across devices. Use navigator.getGamepads() or a library (e.g. gamepad API wrapper) for consistent access.
• Order: Process input → update logic → render. Avoid reading input mid-update.