How to backtest NQ futures for free (step by step)
Backtesting means checking how a trading idea would have performed on past data before you risk money on it. For NQ (the E-mini Nasdaq 100 future) you can do this for free with bar replay: you replay a historical day candle by candle, trade it as if it were live, and record every result.
Manual or automated backtesting?
There are two ways to backtest. Automated backtesting runs code over years of data in seconds. It is fast, but only works for strategies you can write as exact rules, and it is easy to over-fit: tuning the rules until they look perfect on the past and fail in the future.
Manual backtesting with bar replay is slower but closer to real trading. You see the chart forming, you make the decision in real time, and you practise execution, patience and risk management at the same time. For discretionary styles such as liquidity sweeps, fair value gaps or support and resistance, manual replay is often the only honest way to test.
Step 1: write your rules down first
A backtest is only useful if the rules are fixed before you look at the data. Write down, in one page:
- Market and session: for example NQ, New York open, 09:30–11:00 New York time.
- Setup: exactly what must happen before you consider a trade.
- Entry: the trigger, such as a candle close or a break of a level.
- Stop loss: where it goes and why.
- Take profit: a fixed reward-to-risk ratio or a target level.
- Filters: when you do not trade, for example during consolidation or right before news.
- Risk: how many contracts, and your maximum loss per day.
Step 2: replay days you have not seen
Hindsight is the biggest enemy of manual backtesting. If you remember what a day did, you will trade it better than you could live. Two habits fix most of this: pick random days, and hide the date so you cannot connect the chart to news you remember. In NQ Replay, press Random and Hide date, choose a start time a few minutes before your session, and press Play.
Step 3: trade it like it is live
Use the same contract size, stop and target rules you would use with real money. Let the replay run at a speed where you still have time to think, and step bar by bar near your entries. Place the stop loss with the order, not afterwards. If a setup appears that is not in your rules, skip it and write a note.
Step 4: log every trade
For each trade record the date, direction, entry, stop, target, exit, result in points and dollars, and the result in R (the profit or loss divided by the amount you risked). A simulator journal does this automatically and adds commissions and slippage, which matter more than most traders expect on a short-term strategy.
Step 5: judge the results with the right numbers
| Statistic | What it tells you |
|---|---|
| Number of trades | Below about 50 trades, results are mostly noise. Aim for 100+. |
| Win rate | Only meaningful together with average win and average loss. |
| Average R | Your edge per trade in units of risk. Positive after costs is the goal. |
| Profit factor | Gross profit divided by gross loss. Above 1 is profitable; above 1.5 is solid. |
| Max drawdown | The deepest fall from a peak. Decides whether you survive prop firm rules. |
| Results by month | A strategy that only made money in one month is not a strategy yet. |
Common mistakes
- Changing the rules halfway through and keeping the old results.
- Testing only trending days, or only days you remember.
- Ignoring commissions and slippage.
- Stopping after 15 trades because the result looks good.
- Taking a different size or a wider stop in the backtest than you would live.
Done properly, manual backtesting does two jobs at once: it tells you whether an idea has an edge, and it trains you to execute it. That is why many prop firm traders replay dozens of days before paying for an evaluation.
Frequently asked questions
How many trades do I need for a reliable backtest?
As a rule of thumb at least 50 trades, and 100 or more before trusting win rate and profit factor. Fewer trades are mostly noise.
Is manual backtesting as good as automated backtesting?
It answers a different question. Automated tests are fast and exact for rule-based systems; manual replay is better for discretionary setups and also trains your execution.
Should I include commissions in a backtest?
Yes. On short-term NQ and MNQ strategies, commissions and slippage can turn a small edge into a loss.