Skip to main content

Strategies & backtesting

Test the strategy before trusting the signal.

Strategies in Seven A.M. are versioned, parameterised and tested the same way every time — with costs, drawdown and out-of-sample data in plain view.

Backtest report · trend EMA 20/50/200

Illustrative demo — not a real result
Strategy equity (net of fees)No-trade baselineOut-of-sample window
Illustrative equity curve with in-sample and out-of-sample splitIllustrative demo — not a real result. Equity against a flat no-trade baseline; the shaded region is the out-of-sample period.In-sample · 70%Out-of-sample · 30%

Drawdown

Illustrative drawdownIllustrative demo — not a real result. Peak-to-trough decline at each step.
Max drawdown
-5.9%
Simulated fees
3.15 USDT
Fee model
0.10% / side
Slippage model
2 bps
Illustrative demo — not a real result. Generated sample data illustrating the report layout. Backtests are hypothetical, are subject to overfitting, and do not predict future results.

Strategy library

Starting templates.

Templates are starting points for research in the Strategy Lab. None of them is a recommendation, and none has a validated edge: trend, breakout and mean reversion run as shadow forecasters and have not passed the promotion gates.

  • Trend following

    Status: Beta

    EMA 20 / 50 / 200 alignment with configurable confirmation.

  • Donchian breakout

    Status: Beta

    Channel breaks with lookback and volatility filters.

  • Bollinger mean reversion

    Status: Beta

    Band extremes with configurable width and exit rules.

  • No-trade baseline

    Status: Beta

    The control every strategy is measured against.

  • Momentum

    Status: Planned

    Rate-of-change and relative strength across timeframes.

  • Order-book imbalance & regime filter

    Status: Planned

    Planned research models; require additional data feeds.

Methodology

What every backtest includes.

  • Costs

    Exchange-style fees and a slippage model are applied to every simulated fill.

  • Out-of-sample validation

    A held-out window the parameters never saw, reported separately.

  • Drawdown and exposure

    Maximum drawdown, time under water and exposure shown next to returns.

  • Versioning

    Every parameter change creates a new strategy version, so results stay attributable.

  • Same risk rules

    Backtests use the same risk-engine limits as paper trading.

  • Known limits

    Backtests are hypothetical and can overfit. See the risk disclosure.

See it early. Shape it with us.

Join waitlist Open dashboard