Phase 1
Verified Core
- Nine-strategy native MT5 engine
- Central execution layer
- Portfolio risk controls
- Protective stop engine
- Economic-event controls
- Deterministic real-tick validation
- Event-driven state tracking
In Development / Roadmap
What Gunit is being built into
Items on this page are planned or in progress. They are not released features.
In Development / Roadmap
Gunit is being built beyond static automation.
Everything in this section is planned or in progress and is not presented as a released feature. Gunit should improve by measuring strategies independently and allocating risk intelligently — not simply by increasing lot size.
Phase 2
Strategy Intelligence
- Per-strategy profit and drawdown attribution
- Strategy quality scoring
- Dynamic strategy weighting
- Better strategy-combination selection
- Regime-aware strategy activation
- Volatility-aware risk allocation
- Session-aware strategy weighting
Phase 3
Execution Intelligence
- Spread-quality scoring
- Slippage tracking and execution-quality analytics
- Pending-order freshness logic
- Session/liquidity filters
- Broker-profile tuning
- Reduced redundant order-management traffic
Phase 4
Portfolio Optimisation
- Risk-adjusted portfolio optimiser
- Correlation-aware strategy exposure
- Adaptive aggregate-risk allocation
- Drawdown-aware de-risking and recovery
- Walk-forward optimisation
- Out-of-sample validation
- Monte Carlo robustness analysis
Phase 5
Analytics & User Experience
- Per-strategy dashboard
- Balance/equity and drawdown analytics
- Trade attribution
- Risk-at-entry reporting
- Session and market-condition breakdowns
- Exportable reports
- Preset / profile manager
- Safer broker/account compatibility diagnostics
Research philosophy
Measure first. Optimise second.
Gunit development treats signal quality, execution quality and risk allocation as separate problems. Changes are tested against controlled data and consistent settings before they become part of the trading system. The goal is not to maximise one backtest — it is to find improvements that remain useful across different market periods, volatility regimes and execution conditions.
Controlled A/B Testing
Change one meaningful variable at a time.
Out-of-Sample Validation
Optimisation should survive data it was not fitted against.
Risk-Adjusted Improvement
Better return only matters when the additional drawdown and exposure are understood.