A strategy that only works with its exact settings was fitted to the past. The Monte Carlo check (a test with random changes) backtests many slightly changed versions of your strategy and counts how many still hold up.

Plan: Pro (€29 a month). New accounts get 14 days of Pro when they first sign in to the app.

Plan the run

1. Open the review (see Start a robustness review) and scroll to Monte Carlo. On a new review the planner is already open; on one that has run, click Plan a new run.
2. Set Draws: how many changed versions to test, from 8 to 20. More draws give a firmer verdict and take longer.
3. Leave Max change (%) (how far an input may move, 20% by default) and Change probability (how likely each input is to move, 0.5) as they are unless you have a reason.
4. Choose the extra noise: Skipped trades drops 10% or 20% of trades at random (it needs at least 20 trades). Modelled spread adds 1 or 2 pips of spread; the app offers it only for results tested on Darwinex. Sizing inputs such as lot size stay fixed unless you tick Include sizing inputs.
5. Under Inputs to vary, keep Automatic, or pick Select to choose the inputs and their ranges yourself.
6. Click Preview plan to see the draws, then Run N backtests.

Each draw also runs at its own execution delay (0 to 500 ms). For every delay used, the app also re-runs your unchanged strategy, so each draw is compared like for like. That is why 10 draws can mean 14 backtests. They wait in the Queue, and the verdict appears when they are done.

Read the verdict

A draw held up when it kept at least half the unchanged run's net profit and half its trades, with no more than twice its equity drawdown.

• PASS: 80% or more of the draws held up. Conclusion: Pass.
• WARN: 60% or more. Conclusion: Caution.
• FAIL: under 60%. Conclusion: Do not trade.
• INSUFFICIENT: fewer than 8 draws could be judged, or more than a quarter produced no report. Run it again.

Look at the draws

The Draws table shows what each draw changed (for example Stop_Loss 450 → 500), its delay and skipped trades, its net profit, the change against the unchanged run, and Held up or Failed with the reason. View shows one draw in detail, and the note box keeps your comments on it.

Good to know

• The run stops early once the draws still waiting cannot change the verdict. Those show as Not run; they are not failures.
• The same Seed gives the same plan. Re-roll draws a different one.
• Delete removes this attempt from the review; the backtests behind it are kept.
• Robustness › Monte Carlo lists every review's draws, share held up and verdict.

Related guides

• Start a robustness review
• Cross-data check: the same test on another price feed
• Send the strategies that pass to the Portfolio Builder