Methodology

Every strategy begins its life in paper trading mode. Paper trades execute against real market data with simulated fills, allowing us to validate behavior without risking capital.

Only after a strategy demonstrates consistent edge in paper mode -- with honest metrics including spread costs -- is it considered for live deployment. Live trading uses real money on OANDA (forex) and Alpaca (options) accounts.

The dashboard clearly labels every piece of data as either Paper or Live so you always know which environment produced it.

Backtests replay historical data (typically 180 days of H1 candles) to measure how a strategy would have performed. All backtests include realistic spread costs -- we never inflate numbers.

Forward tests (paper trading) run strategies in real-time against live market conditions. This catches issues that backtests miss: slippage, requotes, and market microstructure effects.

We track both side by side. If forward test results diverge significantly from backtest expectations, we investigate before deploying live.

Every strategy follows a strict promotion path:

1. Backtest validation -- must show positive expectancy after spread costs across all deployed pairs/symbols.
2. Paper deployment -- runs on simulated money with real market data. Minimum observation period before promotion.
3. Shadow tracking -- multiple strategy variants run simultaneously on the same trades, measuring which exit rules or parameters would have performed best.
4. Live deployment -- only after paper results match backtest expectations and shadow data confirms the approach.

Parameter changes to the live baseline require a 3-way comparison test (old params vs new params vs control) before deployment.

Risk management is enforced at multiple levels:

• Per-trade risk caps -- forex uses 1-2% of account balance per trade; options uses 20% on a dedicated small account.
• Concurrent position limits -- maximum simultaneous trades prevent overexposure to correlated moves.
• Hard stop losses -- every trade has a pre-calculated stop loss (ATR-based for forex, defined risk for options spreads).
• Bot tagging -- each trading instance uses unique client tags for deconfliction when multiple bots trade the same account.
• API guards -- live order execution requires explicit environment variable authorization.

This dashboard shows real performance data from autonomous trading systems. Nothing is cherry-picked or edited.

We intentionally show:

• Win rates (including losing streaks)
• Profit factors with full spread costs included
• Equity curves showing drawdowns
• Shadow strategy experiments and their outcomes
• Backtest results for strategy validation

Some information is kept private to protect the trading edge and account security:

• Exact account balances and position sizes
• API credentials and account identifiers
• Proprietary strategy parameters and thresholds
• Real-time pending order prices

The telemetry shown here is aggregated and delayed enough to be useful for transparency without compromising security.