Backtesting honestly: what a real report shows

Anyone can produce a backtest that looks profitable. Curve-fit the parameters, ignore costs, test on the one period where the strategy happened to work, and you'll get an equity curve that climbs beautifully. It won't survive contact with a live market. An honest backtest is built to find the flaws, not hide them.

The numbers that matter

A useful report goes well beyond total return:

• Maximum drawdown — the worst peak-to-trough decline. This is what tests whether you can actually hold the strategy.
• Recovery time — how long the strategy stayed underwater. A 20% drawdown that recovers in a week is very different from one that takes a year.
• Profit factor — gross profit divided by gross loss. Below ~1.3, the edge is thin.
• Trade distribution — is the return driven by many small wins, or one lucky trade? Concentration is fragility.

The assumptions that quietly lie

Two settings can turn a losing strategy into a winning-looking one:

1. Zero slippage — assuming you always fill at the exact signal price. Real fills are worse, especially in fast markets.
2. Zero commission — ignoring per-trade costs. For a high-frequency strategy, costs alone can erase the edge.

Model slippage and commission realistically, or the backtest is fiction. We'd rather show you a strategy that's marginal-but-real than one that's spectacular-but-fake.

Parameter sensitivity

The most revealing test: nudge each parameter slightly and see what happens. A robust strategy degrades gracefully. A curve-fit one collapses the moment you move a threshold by a single point. Our reports include a sensitivity table for exactly this reason.

What you get

When you send us a strategy, we return a written report with the equity curve, drawdown analysis, the metrics above, a sensitivity table, and honest commentary on the edge cases we observed. No hype, no guaranteed returns — order one here.