PurrDiction

Overview

Edit this page on GitHub →

The Client-Side Prediction (CSP) system is designed to provide a seamless and responsive multiplayer experience by predicting game states locally on the client. This system is built around a modular architecture that allows for flexible and extensible design, enabling systems that behave as if the game were single-player, while still synchronizing with the server automatically.

Key Components

  1. PredictionManager:
    • Acts as the central "world" for client-side prediction.
    • Manages all predicted entities and systems within the scene it resides.
    • Handles the lifecycle of predicted states, including prediction, reconciliation, and view updates.
  2. PredictedIdentity:
    • Unity components that define the behavior of predicted entities.
    • Created by users to handle specific functionalities, such as movement, physics, or custom logic.
  3. PredictedHierarchy:
    • Provides a prediction compatible version of Unity's Instantiate and Destroy methods.

Design Philosophy

  • Decoupled from Traditional Networking:
    • This system is completely disconnected from the usual NetworkIdentity setup.
    • RPCs (Remote Procedure Calls) are not supported or needed in this architecture, as prediction handles state synchronization naturally.
    • Logic is executed locally on the client, mimicking a single-player experience, while still maintaining consistency with the server.
    • This approach simplifies development, as developers can focus on writing game logic without worrying about networking intricacies.

Benefits

  • Responsive Gameplay: Predictions provide immediate feedback to the player, reducing the perceived latency.
  • Modularity: Systems can be easily added or modified, allowing for flexible and scalable game design.
  • Consistency: Reconciliation ensures that the client’s state aligns with the server’s authoritative state, maintaining a consistent game world.

Limitations

  • Prediction Errors: Incorrect predictions may require corrections, which can occasionally result in visual "snapping" or adjustments.
  • Complexity: While the system simplifies networking, it introduces new challenges in managing predicted states and reconciliation.