Traditional technical analysis relies on "shadows"—lagging indicators like moving averages and RSI that describe what happened without explaining how it happened. Komorebi discards classical time-series analysis entirely, treating the market instead as a dynamical system governed by physical laws.
The engine maps the market into a 4D Computational Fluid Dynamics (CFD) grid. By solving non-linear Partial Differential Equations (PDEs) in real-time, we observe the flow of market energy as a thermodynamic fluid. We measure the Lyapunov Exponent (λ) to identify the exact moment the market transitions from a chaotic, unpredictable state into an actionable, directional "Ignition Zone."
The system is split into two layers: The Macro Layer, where an evolutionary AI (AdaEvolve) calibrates the physics constants across years of historical data, and The Micro Layer, a 100% deterministic execution loop that operates with zero-latency precision.
Before a trade is executed, the market must pass through four strict checkpoints:
• Cold Pressure Building: Detection of structural energy accumulation.
• Mass-Veto Clearance: Ensuring the fluid topology is clear of "phantom" signals.
• Kinetic Ignition: Confirmation of violent gradient shift via the Lyapunov Exponent.
• ABR Fitness: Real-time momentum validation against out-of-sample predictability.
Over a rigorous out-of-sample backtest spanning ~2.8 years, the Sinkhorn-regulated PDE achieved strict statistical significance: 2,301 executions with a 34.75% Win Rate and a massive 27.42x Outlier Multiplier. The engine does not predict; it isolates mathematically guaranteed 27x asymmetric convexity.
Komorebi is built for survival. When the market collapses into high-entropy turbulence, the engine's physics-based laws point-blank refuse to execute.
Komorebi · Thermodynamic Alpha Engine · ψ(x, t)