What results and metrics does a run produce?

A Reamer Research run produces 31 summary metrics, every closed trade, every fill and every order with its outcome, as one schema-versioned JSON document. From Python the binding returns less, the main profit and cost totals and the closed trades, so the full set comes through the C interface.

The 31 summary metrics

• Profit and cost: gross and net profit, with fees, slippage and overnight swap each reported separately.
• Returns and ratios: total return, win rate, profit factor, expected profit per trade, recovery factor, and the Sharpe, Sortino and Calmar ratios.
• Orders: how many were placed, filled, expired, cancelled and rejected.
• Trades: longest winning and losing streaks, and the average, longest and shortest holding times.
• Drawdown: the largest fall in equity, in money and as a fraction.

The records behind them

• Closed trades, each with entry and exit time and price, size, costs, profit and the stop and target it carried.
• Fills, one per execution.
• Orders, every one the strategy placed, with its type, lifetime, outcome, fill price and, if rejected, the reason. Orders still open at the end are listed separately.
• Futures rolls, if you set them up.

The JSON report

The whole result is one JSON file, with a schema document in the kit. It carries a schema version, and within a version fields are only ever added, so your parsing code keeps working on newer releases. The same inputs and seed give the same bytes, so two reports can be compared with a plain file diff.

Limits worth knowing

• No bar-by-bar equity curve. Returns are per closed trade, and the ratios are computed from those. Build a daily curve yourself from the trade records if you need one.
• Annualisation is inferred from the typical gap between trade exits, so a strategy that trades irregularly may not match the convention you expect. You can recompute the ratios from the per-trade returns with your own factor.
• The Python binding returns the main totals and the closed trades, up to 4,096 by default. For the full report, call the C interface.

See How do I know if my backtest is overfit? and What is Reamer Research?