Execution Engine
The production environment owns live strategy decisions, exact account binding, risk limits, stale-data protection, order-state reconciliation, operator controls, and fail-safe behavior.
VORAX is the internal market infrastructure of Kim Capital Holdings. It supports research, execution governance, evidence capture, strategy testing, and live virtual validation—without allowing the research layer to quietly become financial authority.
VORAX is not presented as a black-box promise of effortless returns. It is infrastructure: deterministic rules, risk controls, data quality, experiment lineage, live observation, and explicit human authority.
The production environment owns live strategy decisions, exact account binding, risk limits, stale-data protection, order-state reconciliation, operator controls, and fail-safe behavior.
Forge tests multiple hypotheses against shared market facts, immutable datasets, realistic costs, time-window discovery, stress testing, and reproducible evidence packages.
Shadow Desk watches completed live-market bars, produces virtual opportunities and trades, tracks current regime, and compares recent behavior with historical research evidence.
Systems should increase clarity, repeatability, and control—not conceal uncertainty behind automation.
Every layer has a narrow contract. Data creates features. Features create reason-coded proposals. Simulation creates evidence. Evidence creates a review decision. No research output promotes itself.
Canonical completed bars, timestamps, volume, quality checks, and immutable identity.
VWAP, ATR, trend, structure, volume, session state, and market regime.
Deterministic hypotheses create proposals or explicit reason-coded no-signal decisions.
Spread, slippage, delay, stops, targets, risk sizing, and ambiguous-bar policy.
Trade ledgers, reports, robustness, holdouts, Shadow behavior, and human-only progression.
The discretionary swing process and the automated systems layer can challenge, confirm, or diversify one another. The principal still owns the final mandate, portfolio context, and decision authority.
Historical replay, time discovery, parameter stability, and regime decomposition expose where an idea works—and where it only appeared to work.
Shadow Desk measures how systems react to fresh completed bars without turning research code into a path to live orders.
Production authority, research authority, operator authority, and evidence are separated so that no favorable backtest can bypass risk and review.