Lessons › Advanced track › Manual backtesting

World 10 · Advanced track · lesson 7 · level 4

Manual backtesting

Before trusting an idea, test it on lots of old charts with fixed rules and write down every result.

In one line

You think your setup works. What would 40 honest trades on old charts say?

Explained simply

A manual backtest is like practising a recipe many times before a big dinner and writing down how each attempt turned out. You follow the same steps every time, even when you suspect it will flop. Only then do you know how good the recipe really is.

The lesson

A backtest replays history to see how a rule would have performed. In a manual backtest you move forward one candle at a time with the future hidden, take every trade that fits the written rules, and record the entry, stop, exit and result in R. Testing at least 30 to 50 trades across different market conditions, such as trends, ranges and sharp falls, gives a rough first picture, not proof. Skipping signals that look bad in hindsight quietly inflates the results.

A worked example

Illustrative example: your breakout rule triggers 40 times over three years of daily charts. You log 16 wins averaging +1.8R and 24 losses of -1R. The total is 16 x 1.8 - 24 x 1 = 4.8R, so the expectancy is 4.8 / 40 = 0.12R per trade before costs, and the win rate is 16 x 100 / 40 = 40%. If you had quietly skipped 4 ugly-looking losing signals, the total would have looked like 4.8 + 4 = 8.8R, which is far too rosy.

The same idea at four levels

1. Beginner. A backtest checks how your rule would have done on old charts.
2. Foundation. Move one candle at a time, keep the future hidden, and take every trade the rule gives.
3. Intermediate. Log the date, setup, entry, stop, exit, result in R and notes for every trigger, including the ones you don't like.
4. Advanced. Test across trending, ranging and crash periods, because a rule tested only in one kind of market can look far better than it really is.
5. Expert. Treat 30 to 50 trades as a first look: work out the expectancy and the worst losing streak, then confirm on fresh data before risking anything real.

Mistakes to avoid

• Skipping signals that look bad in hindsight.
• Testing only on a period where you already know the idea worked.
• Peeking at the next candles before deciding whether the rule triggered.

Goal of this lesson: Test a rule-based setup on past charts by hand, recording every trade honestly.