Backtesting
Test Ten Years of Your Strategy in Minutes, Not Days
Sterling Partners runs tick-level backtests against historical exchange data from 2014, with realistic slippage models and broker commission schedules built in.
What a Sterling Partners Backtest Actually Tests
Most backtesting tools apply bar-level close prices and assume perfect fills. The Sterling Partners engine replays tick-by-tick order-book events and simulates queue position to estimate realistic fill probability and slippage. Commission schedules from 12 supported brokers are pre-loaded and applied per-trade, so your net-of-costs P&L is what you see in results — not a gross figure that falls apart when you go live. Walk-forward optimisation runs are parallelised across up to 16 CPU cores, so a 10-year monthly walk-forward on a complex multi-leg strategy completes in under eight minutes.
Backtest Output: What You Receive
Every completed backtest generates a structured report you can interrogate and export.
Trade-by-trade log
Every simulated trade: entry timestamp, fill price, size, exit, P&L, slippage estimate, and the specific rule condition that triggered it. Exportable to CSV for external analysis in R or Python.
Equity curve and drawdowns
Interactive equity curve with max drawdown periods highlighted. Click any drawdown trough to view the underlying trade log for that period — so you understand what market condition caused the underperformance, not just that it happened.
Parameter sensitivity
Heat-map of strategy performance across a grid of parameter values — spot the difference between a robust parameter region and an over-fitted peak. Prevents you from selecting settings that look exceptional in-sample but collapse out-of-sample.
Backtests Are Evidence, Not Guarantees
A backtest is a model of the past, and past market conditions do not repeat identically. Results from historical simulation will always overstate the quality of a strategy to some degree — even tick-level models cannot perfectly replicate live execution in a fast-moving market. Walk-forward testing and out-of-sample validation reduce but do not eliminate this gap. Treat any backtest result as a lower bound on what you need to see before committing live capital, not as a forecast of future performance.
Run Your Strategy's First Backtest Today
Upload your rules or connect the Python SDK and run a full 10-year backtest free during your trial period.
Start your trial