AI Player Behavior Analysis & Difficulty Adjustment

You are a game data scientist. Design an AI player behavior analysis and dynamic difficulty system: Game type: [INSERT - mobile, live service, single-player, multiplayer] Engine: [INSERT - Unity, Unreal, custom] Data backend: [INSERT - local, cloud analytics] Difficulty style: [INSERT - rubber banding, dynamic, procedural] Cover the full pipeline: 1. Telemetry design - events to track (deaths, time in level, resources used, choices, session length), data schema, sampling, privacy compliance. 2. Player profiling - skill estimation (Elo/Glicko), playstyle clustering, engagement metrics, churn risk prediction, frustration signals. 3. Difficulty model - difficulty parameters, player vs challenge matching, frustration/flow/boredom zones, difficulty update frequency. 4. Real-time adjustment - dynamic enemy AI scaling, resource spawn tuning, level selection, assist features, pacing changes. 5. Offline analysis - cohort analysis, funnel analysis, A/B test design, balancing difficulty curves from real data. 6. Fairness and anti-cheat - detect boosting/exploits, prevent trivializing content, preserve challenge for skilled players, avoid pay-to-win feel. 7. Player agency - transparency vs invisibility of DDA, player-selectable difficulty, feedback when difficulty changes, accessibility options.