How to backtest a trading strategy for beginners

What you are doing

A backtest asks a plain question: if I had followed these exact rules on past prices, which trades would I have taken, and how did each one end? You write the rules, a tool applies them to old data, and you read the record it gives back.

It does not predict anything. Simulated results are not actual trading. They show what the rules would have done on the history you tested.

Rule

One instruction with a single reading, such as when to enter, where the stop goes or when to exit.

Stop

The price at which a trade is closed at a loss.

Target

The price at which a trade is closed to take a gain.

Bar

A summary of a stretch of time: the open, the high, the low and the close.

Step 1: Pick one simple idea

Choose one setup you can describe in two sentences. A previous-day-high break, an opening-range break and a moving-average cross are common starting points. Skip the idea that needs a long list of conditions. Each condition removes trades and adds ways to be wrong.

Start with one rule you can say in two sentences. Add nothing until you have read every trade it produced.

Step 2: Write it so it cannot be misread

Name the market, the time zone, the entry, the stop, the target and what happens if neither is hit. Then ask of each sentence: could someone read this another way? If so, tighten it.

Step 3: Choose the data and the market

Use bars fine enough for the rules. A rule about the first minutes after the open needs minute bars. Pick one market to begin with. On our data, NQ and ES run from June 2, 2016 through August 27, 2026.

Step 4: Set the costs

Add commission and slippage before you run anything. Commissions and fees vary by broker; a backtest should use your own. A result with no costs is one no account could have had.

Step 5: Run it and read every trade

Run the rules, then open the trades. Check a few against the chart. Look at how many trades there are, and at how they spread across the years. Read the figures next to the count, because a figure from a handful of trades says little.

Step 6: Hold some months back, then change one thing

Set aside part of the history before you look at results, and test the finished rules on it once. On our service about 30% of months, spread across every year, are held back for each strategy.

Then change one rule at a time and run again. If you change three things at once, you cannot tell which one moved the result.

Where to practise

Our Trade Lab is a free bar-by-bar replay on real NQ, ES and GC data, with paper orders and a journal. It needs an account with a verified email. For a full run, describe the rules in plain language and we test them. A version is $19.