Backtests
EA trading costs: stress the net result
Start with costs already included in net P/L. Add only the extra stress amount, and state whether execution was rerun or the existing trade list was merely adjusted.
Build the cost ledger first
Reconcile gross trading outcomes, commission, swaps, fees and ending balance. Identify whether spread is embedded in fill prices and whether an export column is already net of other charges. Applying a percentage increase to undefined costs is not reproducible. Cash deposits and withdrawals must remain outside the trading profit calculation.
Reruns and deductions answer different questions
A native rerun can change whether an order opens, how it fills and when a stop triggers. A post-hoc simulation keeps the original trades and subtracts modeled incremental charges. It is useful for sensitivity, but cannot establish the equity path or trading decisions of a changed execution environment. Label these methods separately on every result graphic.
Challenge cost-sensitive intervals
Inspect rollover, weekly opening, thin sessions and unusual spread widening. Apply volume and holding-period dependence when appropriate rather than one arbitrary fee per trade. Inspect gaps and adverse fills separately; a smooth 50% cost uplift is not a model of every discontinuous execution failure. Set stress criteria before comparing candidates.
| Scenario | Extra cost | Net P/L |
|---|---|---|
| Baseline | EUR 0 | EUR 1,200 |
| +50% on EUR 1,000 included costs | EUR 500 | EUR 700 |
| +100% on those costs | EUR 1,000 | EUR 200 |
Illustrative examples and review assumptions; not measured TradeBerger EA results.
Checks to make before deciding
- Export the original MT5 report, journal and deal list from the same completed run. Save the report before another test overwrites it and verify its embedded settings.
- Declare whether trade count means exit deals or completed positions. Handle partial closes, scale-ins, reversals and multiple tickets without silently double-counting strategy opportunities.
- Reconcile starting balance plus realized profit, commission, fees, swaps and cash movements to ending balance. Explain rounding tolerances and any residual discrepancy.
- Check cost signs and whether exported profit already includes costs. Do not subtract the same commission or swap twice when creating monthly performance or stress simulations.
- Build a baseline cost ledger separating spread, commission, fees, swaps and slippage assumptions. Confirm which components are already embedded in exported profit.
- Predeclare moderate and severe cost scenarios, for example +25%, +50% and +100% of defined included costs. These are test choices, not predictions of broker conditions.
- Prefer native MT5 reruns when changed spread, delays or sessions can affect entries and exits. Save the new settings and inspect altered trade counts and fills.
- For post-hoc stress, subtract only incremental costs from the existing net series. Label the method as modeled and state that trade selection and execution remain unchanged.
FAQ
Does +50% cost stress mean +50% spread?
Only if you explicitly define that scenario. A combined cost budget, commission uplift and spread change are different assumptions.
Can I claim native equity DD after deducting fees?
Not from a closed-trade adjustment alone. Reconstruct suitable marked-to-market data or explain that only closed P/L sensitivity was modeled.