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 resultDrawdown
- Max drawdown
- -5.9%
- Simulated fees
- 3.15 USDT
- Fee model
- 0.10% / side
- Slippage model
- 2 bps
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: BetaEMA 20 / 50 / 200 alignment with configurable confirmation.
Donchian breakout
Status: BetaChannel breaks with lookback and volatility filters.
Bollinger mean reversion
Status: BetaBand extremes with configurable width and exit rules.
No-trade baseline
Status: BetaThe control every strategy is measured against.
Momentum
Status: PlannedRate-of-change and relative strength across timeframes.
Order-book imbalance & regime filter
Status: PlannedPlanned 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.