Strategies & Robots
The five built-in trading strategies from the Aegis quant engine, each with an ATR stop, a 2:1 target, and market-regime detection.
Built-in strategies
Aegis ships five built-in strategies, computed by a dedicated Rust quant engine rather than in your browser. Each one turns market data into entry, stop, and target levels, and every strategy shares the same risk scaffolding: an ATR(14, 2x) stop and a 2:1 reward-to-risk target, combined with market-regime detection (trending up, trending down, ranging, breakout, and high/low volatility) so the robot only acts when conditions fit its edge.
Automation starts DISABLED. A strategy generates signals only after you explicitly opt in, and you can halt it at any time — see Risk Management.
| Strategy | Idea | Best regime |
|---|---|---|
| Momentum | Ride the prevailing move — buy strength, sell weakness — while momentum indicators stay aligned. | Trending up or down |
| Breakout | Enter as price clears a defined level or range boundary, targeting the expansion that follows. | Breakout, rising volatility |
| Trend Following | Stay with an established direction until the trend structure breaks. | Sustained trending |
| Mean Reversion | Fade stretched moves back toward a statistical mean when price is over-extended. | Ranging, low volatility |
| Ensemble | Weighted vote of the other four strategies, blending their signals into one decision. | Mixed or uncertain |
How signals are formed
The quant engine draws on 10 indicators — ADX, ATR, Bollinger Bands, EMA, MACD, Momentum, RSI, SMA, Volume, and VWAP — to classify the current regime and produce a signal. Each signal carries the entry price, the ATR-based stop, and the 2:1 target, which flow through to the order layer.
Because the stop and target are attached to every signal, orders placed from a strategy are risk-defined from the start. See Orders & Executions for how those levels become live orders, and Backtesting to validate a strategy over history before enabling it.
Machine learning
A Python training pipeline exports ONNX models that the quant engine can run. Live, on-chart inference wired into signal generation is In development — today's signals come from the rule-based strategies and indicators above.