ZenWave engines are built using AI-accelerated Walk-Forward Analysis — a structured framework that prioritises out-of-sample validation over in-sample curve-fitting.
ZenWave uses a dual-AI research pipeline. AI is employed as a co-researcher — never as a black-box decision-maker. The system leverages:
Generates strategy candidates, assists with feature exploration, and accelerates the hypothesis-to-code cycle.
Independently reviews Walk-Forward results, checks for overfitting signals, and validates procedural integrity.
Every critical decision — engine selection, parameter acceptance, risk allocation — is made by the human researcher. AI accelerates discovery; it does not govern deployment.
Walk-Forward Analysis is the gold standard for validating algorithmic trading strategies. It divides historical data into sequential in-sample (IS) and out-of-sample (OOS) periods:
ZenWave engines use multi-slice anchored WFA. Each engine has multiple slices (A1–A4, B1–B5) to assess parameter stability and regime resilience.
The entire ZenWave system targets a single instrument: USD/JPY. This is not a limitation — it is a deliberate design choice grounded in five structural advantages:
Focusing on a single pair removes cross-pair noise, avoids correlation overlap, and allows every engine component to be tuned for the specific micro-structure of one market.
All risk boundaries are fixed before deployment and never adjusted in response to live drawdowns. The framework includes:
Total portfolio risk capped at 0.50% per trade cycle, split equally across active engines.
A and B run independently. A drawdown in one engine does not trigger adjustments in the other.
No discretionary intervention. Every trade is generated, managed, and closed by the engine.
ZenWave deploys a multi-horizon portfolio rather than a single strategy. Each engine targets a different time-horizon:
Multi-horizon diversification reduces decay risk, broadens regime coverage, and mirrors institutional portfolio design.
Everything published — every claim, every chart, every performance statistic — is backed by reproducible, auditable evidence. ZenWave follows four non-negotiable principles:
Performance is measured only on out-of-sample data. In-sample numbers are never reported as expected performance.
WFA slices, parameter ranges, and OOS equity curves are published on YouTube and in quantitative reports.
Live results are tracked on cTrader Copy — publicly visible balance, equity, and trade history.
An independent AI reviews validation procedures and checks for procedural errors or overfitting signals.