Advanced Behavioral Simulation System for WRobot

Overview

This combat class represents an experimental AI-driven Warrior Tank system designed for WRobot.
It combines predictive movement, adaptive hysteresis, and context-aware decision making to simulate highly realistic player behavior during combat.

The system is built around multiple synchronized threads controlling both rotation and movement intelligence independently, allowing it to react smoothly to complex combat situations and environmental changes.

Core Features

• 🧠 Adaptive AI Movement Engine
  Dynamically positions the player at the threat centroid, repositions based on enemy spread, and adjusts movement smoothness according to FPS, ping, and encounter type (trash, elite, boss).

• ⚙️ Predictive Enemy Tracking
  Utilizes Kalman-based filtering to estimate enemy movement vectors and accelerations, improving facing and threat control accuracy.

• ⚔️ Context-Aware Combat Positioning
  Detects isolated targets, executes Charge or Intercept where appropriate, and uses fallback movement logic when gap closers are on cooldown.

• 🧩 Threaded Coordination System
  Rotation and movement logic run in separate threads, synchronized through custom locks and adaptive state arbitration layers.

• 🧱 Adaptive Hysteresis System
  Dynamically recalibrates movement thresholds based on performance metrics, preventing excessive micro-adjustments and improving realism under unstable frame rates or network latency.

• 🧭 Momentum-Based Movement Model
  Simulates inertia and directional smoothing for more human-like motion transitions.

Entropy & Timing Source

This system uses a high-quality entropy source derived from the operating system scheduler and runtime timing variations.
Key properties:

• The entropy is effectively CPU-free (it leverages natural scheduling jitter, timer granularities and other OS timing artifacts rather than expensive cryptographic operations).

• The behavior produced by this entropy source is inherently non-deterministic and differs across machines — even identically configured systems will exhibit small, meaningful timing differences.

• These timing differences are intentionally used to produce microvariation in movement and decision timing so that runs are not perfectly repeatable.

Note: the entropy mechanism is not a cryptographic RNG; it is a lightweight, practical source of behavioral randomness derived from normal OS operation and scheduling jitter.

Performance and Behavior

This class constantly self-adjusts based on:

• Current FPS and ping

• Number of enemies

• Player health

• Threat status and boss detection

It strives to behave differently under varying system conditions — introducing slight desynchronization across machines.
This creates unique microbehavior signatures, reducing the chance of pattern-based analysis producing identical traces.

⚠️ Important Notice — BETA Version

This system is currently in BETA and contains a carefully tuned internal race condition design.
These controlled timing variances serve as a behavioral anti-analysis layer, ensuring that no two executions behave identically — even on the same computer.

However, due to this design, the AI may sometimes:

• Attempt to move in unintended directions

• Overreact to sudden environmental or combat changes

• Momentarily “break free” or exhibit erratic behavior (“self-destructive impulses”)

➡️ Manual supervision is absolutely required.
Never run this system unattended — it is intended for research and supervised use only.

Known Risks & Stability Notes

• The system uses multiple threads and many fine-grained locks; while designed to be thread-safe, deadlocks or race outcomes can still occur under pathological conditions.

• In extreme cases, the character may attempt risky movement (e.g., run directly into environmental hazards or pathing pitfalls).

• Operator oversight is mandatory — monitor behavior and be ready to interrupt or take manual control.

🧩 Internal Architecture

Main (ICustomClass)
│
├── Thread #1 → RunRotation()
│               - Handles combat, buffs, targeting
│
├── Thread #2 → RunMovementAI()
│               - Handles positioning, facing, LoS, gap closer logic
│
└── Sub-thread → PathfindingOptimizer
                 - Calculates safe spots and escape routes

WarriorTankRotation
│
├── TankMovementAI
│   ├── EnemyMovementPredictor  ← Kalman filter prediction
│   ├── AdaptiveHysteresis      ← FPS / Ping adaptive smoothing
│   ├── PathfindingOptimizer    ← Separate thread for safe pathfinding
│   ├── IsolatedTargetHandler   ← Charge / Intercept decision system
│   ├── ExponentialSmoothing    ← Smooth motion and rotation
│   ├── PerformanceMetrics      ← Internal benchmark & debugging
│   └── DangerState             ← Environment threat evaluation
│
├── TurnCommandArbiter          ← Rotational control priority
│
└── WRobotCoordinator           ← Manages shared movement ownership

Combat Start
   │
   ├─ Is manually controlled? → Pause AI
   │
   ├─ Target in range?
   │    ├─ <5y → Maintain melee
   │    ├─ 8–25y → Use Charge/Intercept
   │    ├─ 5–8y → Fallback to charge range
   │    ├─ >25y → Direct approach
   │
   ├─ Isolated from group? → Smart Gap Closer logic
   │
   └─ Danger detected?
        ├─ Compute escape path (PathfindingOptimizer)
        └─ Defensive rotation & movement priority

🧾 Internal Evaluation

Total Score: 🧩 9.3 / 10

🧰 Requirements

• WRobot for WoW 3.3.5a (WotLK)

• .NET Framework 4.7.2+

• Protection Warrior

💬 Final Words

This FightClass merges combat logic, movement prediction, and manual input detection into a smooth, human-like tank AI.
Perfect for soloing, grinding, or dungeons — safe, adaptive, and powerful.

🧪 Status: BETA
💻 Author: Calaude
🧠 Version: v5.2 (DEBUG + LoS)
🗂️ Build: Advanced_WAR_TANK_192_Optimized