Inputs
- Live and historical market/price data for the configured instruments
- Candidate strategy and model definitions
- Closed and open trade history
- Configured risk limits — per-model position caps, max open trades, drawdown thresholds
Workflow
Market Data Ingestion → Strategy Discovery → Strategy Directory /
Intelligence → Signal Generation → Backtesting → Strategy
Selection → Execution Automation → Risk / Drawdown Monitoring →
Performance Analytics → Strategy Monitoring.
Automation
- Candidate trade generation runs on a continuous loop against live market data
- Strategy comparison recalculates from the underlying closed-trade evidence on every query, not a cached report
- Execution and position-limit gating are enforced automatically, not by manual sign-off
- Drawdown and scaling-defect monitoring flags anomalies without waiting for a scheduled review
Outputs
- A ranked strategy directory with live-recalculated evidence
- Executed and monitored open positions with an audit trail
- Performance and drawdown analytics per strategy and per model
- Anomaly and scaling-defect alerts
Intelligence / data layer
Every strategy comparison is computed directly from closed-trade history rather
than a periodic snapshot, the Strategy Directory below queries the live
database on every request, so the ranking always reflects the current evidence,
not last week's export.
Integrations
- Market data and price-collection services feeding live signal generation
- SQL Server trade database as the system of record for open and closed trades
- Analytics dashboards over the same data for performance and drawdown review
Measurement
- Net return, win rate and profit factor per strategy, recalculated live
- Open and closed trade counts per model and per strategy family
- Drawdown against configured risk limits
- Query and execution latency of the underlying data pipeline
Feedback loop
Live execution and monitoring data feed straight back into strategy comparison
and selection, so underperforming or misbehaving strategies are demoted by
evidence rather than by manual review, the same discipline that surfaced
and fixed a real scaling defect in this platform's own trade-creation pipeline
during development.