Backtest one risk rule in a single evening

Most traders test their entries for months and their risk rules for nothing — the risk rules get tested live, with real money, one expensive lesson at a time. It does not have to work that way. Any risk rule can be checked against history in one evening, with a spreadsheet and a bit of honesty. Here is the exact process.

Pick one rule — only one

A risk rule is a sentence specific enough to be graded true or false. Weak: "I should manage risk better." Strong:

• "Never risk more than 1% of the account on one trade."
• "Every stop sits beyond the most recent swing, with a buffer."
• "No new position if used margin exceeds 20% of equity."

Take one sentence. An evening is not enough for five rules, and testing them together muddies which one actually helped.

Collect 40 setups from history

Open your chart, pick the pair and timeframe you actually trade, and scroll back a few months. Then walk forward bar by bar — without peeking — and mark every place your rule would have applied. For each one, write a single row in a spreadsheet:

Column four is where most people cheat — stops are always beautiful in hindsight. Force yourself to answer: from this bar, knowing only what was visible then, where would the stop have gone? Measure it in pips and write it down before moving forward.

Score it in R, not in money

R means the amount you risked; a win of 2R made twice what it lost at 1R. Working in R keeps the test about decisions, not account size. When the 40 rows are in, the arithmetic is two lines:

• Win rate = winners ÷ total.
• Expectancy per trade = (win rate × average win in R) − (loss rate × 1).

Say 16 of 40 trades won (40%) and the average winner made 2R:

(0.40 × 2.0) − (0.60 × 1.0) = 0.8 − 0.6 = +0.2R per trade

Positive, but modest — which is exactly the kind of truth this process exists to reveal. At 40 trades, +0.2R means the rule roughly pays its own way; it deserves more data, not a levered-up account.

The honesty pass

Three traps ruin otherwise honest backtests, and all three are avoidable:

• Convenient samples. If you notice yourself skipping the ugly, choppy stretches, count them anyway — that is where the rule will actually live.
• Event blindness. Replays that ignore the calendar produce fantasy stops. Mark the days with major releases — the economic calendar shows them — and note how the rule behaved around them before deciding it is robust.
• Sizing amnesia. The test tells you R outcomes; your account cares about dollars. Convert the rule to lots for your real account with the position size calculator so the evening's result maps onto Monday morning.

Paper does not lie about the arithmetic. Only the sample can lie — and 40 honest rows expose a bad rule for the price of one evening, instead of the price of an account.