Guide
How Many Trades Do You Need?
There is no magic number, but the smaller the sample, the less confidence you should place in the observed win rate, drawdown or final account result.
The short answer
Use 20 trades as an early check, 50 trades as a rough checkpoint and 100 or more trades as a more useful review sample. For active systems, 200+ trades usually gives a stronger view.
The exact number depends on trade frequency, market type, strategy style and how much variation the system naturally has.
A practical answer by decision type
The number of trades you need depends on the decision you are trying to make. A small sample can help you find process errors, but it should not carry the same weight as a larger performance review.
| Decision | Minimum useful sample | Better sample |
|---|---|---|
| Check whether rules are clear | 20 trades | 50 trades |
| Review early win rate and losing streaks | 50 trades | 100 trades |
| Estimate expectancy more seriously | 100 trades | 200+ trades |
| Compare market regimes | Multiple 100-trade samples | Separate samples by condition |
Trade count checkpoints
Use these checkpoints as a practical review framework, not as hard statistical rules. The higher the variance of the system, the more trades you usually need.
| Trade count | What it can show | What it cannot prove |
|---|---|---|
| 20 trades | Execution quality and obvious rule issues | True win rate or long-term expectancy |
| 50 trades | Early pattern of win rate, drawdown and streaks | Whether a strategy is definitely good or bad |
| 100 trades | A more useful review of expectancy and variance | Certainty about future performance |
| 200+ trades | Stronger evidence for active systems | Protection from changing market conditions |
What 20 trades can tell you
Twenty trades can reveal obvious execution problems, rule confusion or whether the setup is being followed correctly. It is not enough to prove the true win rate of a strategy.
At this stage, focus more on process quality than final profit or loss.
What 50 trades can tell you
Fifty trades gives a more useful sample, but it can still be distorted by one good or bad sequence. A 50-trade review should include win rate, expectancy, drawdown, losing streaks and whether the results match the strategy logic.
Why 100 trades is a better checkpoint
One hundred trades gives the edge more room to show up. It still does not guarantee certainty, but it is much better than judging from one week of results.
Run 100-trade samples in the trading probability simulator to see how different the path can look even with the same assumptions.
How to review a 100-trade sample
A useful review is not only a profit/loss check. Review the full shape of the sample so you can separate strategy quality from normal variance and execution drift.
- Compare observed win rate with the expected win rate.
- Calculate expectancy using average winner and average loser.
- Review maximum losing streak and maximum drawdown.
- Check whether losses stayed within the planned risk.
- Mark any trades that broke the rules before judging the system.
Keep the sample consistent
A 100-trade sample is only useful if the trades belong to the same strategy idea. If entries, exits, market, timeframe or risk rules keep changing, the sample becomes a mix of different systems rather than evidence about one system.
When the setup changes materially, start a new sample or tag the trades separately. This makes the review less flattering, but much more useful.
When to stop judging and keep collecting data
If the sample is small and the rules were followed, avoid overreacting. Keep collecting data unless the strategy shows a clear structural problem, such as losses much larger than planned or setups that no longer match the original logic.
If the sample is large and expectancy, drawdown and execution are all poor, it may be time to pause and review.
Use simulations while the real sample grows
If you do not have enough trades yet, use the simulator to understand what a normal range of outcomes could look like under your current assumptions. This does not replace real trades, but it helps prevent overreacting to a short sequence.
For example, run several 100-trade samples with the same win rate and reward/risk. If one path looks smooth and another includes a deep drawdown, that difference is a reminder that one live sample is only one possible path.
Frequently asked questions
Is 100 trades enough to judge a trading strategy?
It is a useful checkpoint, but not a guarantee. It is much better than 10 or 20 trades, but more data is usually better.
Do I always need 200 trades?
No. Two hundred trades can give stronger evidence, but the right sample depends on trade frequency, strategy type and how much risk you take while collecting data.
Can I judge a strategy after one bad week?
Usually no. A bad week can be normal variance, especially for strategies with lower trade frequency.
Should I count only live trades?
Live trades are valuable because they include execution and psychology. Backtests are useful too, but they should be treated separately from live performance.
What should I review besides win rate?
Review average win, average loss, expectancy, drawdown, losing streaks, rule adherence and whether market conditions still fit the strategy.
Is 50 trades enough for a high-frequency strategy?
It can be an early checkpoint, but high-frequency strategies usually need more data because small edge differences and costs matter a lot.
Should I mix different strategy setups in one sample?
Usually no. Mixing unrelated setups can hide which idea is actually working or failing. Track distinct setups separately when possible.
Should I restart the sample after changing rules?
Yes, if the change is material. A new entry rule, exit rule, market or risk model can change the system enough that the old sample should be separated.