Describe your rules the way you'd explain them to a friend. Get an honest answer on whether they hold up — not just a pretty equity curve.
Pilot — by requestMetaTrader's tester is excellent at what it does. Today our tool covers gold, 1–3 years of minute data, fixed stop-loss / take-profit rules, and costs based on the fee schedule you provide. Items marked planned are not available yet.
| MT4 / MT5 Strategy Tester | ForwardProof | |
|---|---|---|
| Data & speed | Strong. Tick data, fast multi-core optimisation. | Gold, 1–3 years of minute data; not built for tick-level HFT. |
| How you define rules | Write an Expert Advisor in MQL4/MQL5. | Plain English. We translate it into code and confirm the logic with you. Currently fixed stop-loss / take-profit rules. |
| Luck vs edge | Shows profit on the tested data; optimisation can overfit. | Planned: out-of-sample hold-out, walk-forward testing and statistical significance. |
| Costs | Configurable, but easy to leave unrealistic. | Costs deducted based on the fee schedule you provide. |
| Loss diagnostics | Basic trade list and report. | Planned: MFE/MAE per loser, false-breakout rate, max losing streak. |
| Conclusion | Numbers — interpretation is up to you. | Planned: a plain statement of whether the result met or did not meet the pre-agreed criteria, and why. |
| Price | Free with MetaTrader. | Coming soon. |
Available now: costs and fixed stop-loss / take-profit results. Cards marked planned are not available yet.
Part of the data is locked away during design and only used once, to check the result wasn't fitted.
The strategy is re-tested across rolling windows to see if it survives different market regimes.
How likely is this result by pure chance? If luck explains it, we say so.
Costs are deducted based on the fee schedule you provide, so the break-even reflects your account.
Where losers come from, and how often breakouts fail on your market.
Met the pre-agreed criteria? A backtest alone is not enough — a paper forward test on live data comes next.
We run gold backtests (1–3 years of minute data, fixed stop-loss / take-profit, costs from the fee schedule you provide) for a small number of pilot users. To request one, email aigoldlab@mail.grokbot.com with the subject “Backtest request” and describe your strategy rules in plain English (market, timeframe, entry, stop-loss / take-profit, position sizing, and your fee schedule).