Minecraft anticheat guide
Machine Learning for Aim and Click Detection
See how FairPlay aim and click models fit alongside deterministic checks and staff review.
ML is an additional signal
Machine learning should not be treated as magic. In FairPlay, ML is best understood as an additional confidence layer beside explicit checks for aim, click timing, rotation artifacts, and packet context.
The updated FairPlay description includes model registry, training commands, window recording, model enable/disable/reload, and separate aim and click model pipelines.
Why training workflow matters
Without training workflow, ML becomes a black box. FairPlay exposes data and model workflows so admins can manage windows, confidence, datasets, compilation, and reload behavior more intentionally.
How to use ML safely
Start with ML alerts and confidence output. Compare against explicit check evidence. Avoid instant severe punishments until the server has enough local data. Use ML as part of a layered decision, not the only reason for action.
Model lifecycle and training are distinct
FairPlay includes separate AimBot and AutoClicker machine-learning checks, feature extractors, model services, and training tools. MainConfig.auto-update-models replaces installed models with bundled models on startup.
Machine-learning optimization settings are for training and do not change detection. The presence of a model or training option does not provide an accuracy measure for your server; review alerts alongside other check details.
FAQ
Does FairPlay include ML?
Yes. FairPlay includes aim and click machine-learning workflows, model management, training data, and confidence tooling.
Should ML instantly ban players?
No. Treat ML as one confidence signal and tune it against explicit checks and production data first.