The Settings That Wiped the Account in Our Own Tests, and How Far the Default Sits from Them

EA | Setting it ships with | Where the test broke | What happened

Overdrive Usdjpy Aggressive | RiskPercent 3 | RiskPercent above 7 | account wiped, -99%

Audcad Peace Mean Reversion (averaging grid) | risk multiplier x2.5 at the top preset | multiplier x3 | -99.7%

- Overdrive Usdjpy Aggressive: 7 / 3 = 2.3 times the default.

- Audcad Peace Mean Reversion: 3 / 2.5 = 1.2 times the top preset. One more notch on the multiplier was the difference between +725% and -99.7% in our backtests. That is why the top preset is capped at x2.5.

Neither margin is large. "I will just double it" stays inside the first and goes straight through the second.

- Overdrive Usdjpy Aggressive at its default of 3: profit factor 1.33 and a maximum equity drawdown of 29.37% over 877 trades (2014 to mid-2026).

- Audcad Peace Mean Reversion at the top preset: +725%, profit factor 1.53, maximum margin drawdown 36.2%, relative drawdown 76.0% (2015 to 2024).

A wiped account is the end of a slope, not a cliff that appears from nowhere. The drawdown is already deep one step before it.

On the grid EA the lowest preset halves the lot but keeps the same dollar basket stop. Its result: +32%, profit factor 1.40, maximum margin drawdown 30.4%. The preset above it: +313%, profit factor 1.74, maximum margin drawdown 13.7% (relative 30.8%). The lowest preset had the deeper percentage drawdown of the two. We print that in the description because a buyer would otherwise assume the opposite.

- These are backtest outcomes on historical data. The real break point on a live account can come at a lower setting, because the future can contain a worse stretch than the tested period did.

- A break point found by testing is only as good as the period tested.

- Two EAs is the whole sample. For our other listed EAs we have not published a break point, and that is a gap on our side, not evidence that they have none.

- Ask any seller, us included: at what setting did the backtest fail? If the answer is that it never fails, the test was not pushed far enough.

- Find it yourself. Run the EA at the default, then at double and at triple the risk, over the full period. Note the maximum equity drawdown of each run. The setting where the curve ends is your real upper bound, and a sensible working setting is well below half of it.

- Treat a high default as a choice the seller made for the headline number. You are free to make a different one.

The two EAs in this article: Overdrive Usdjpy Aggressive https://www.mql5.com/en/market/product/182367 · Audcad Peace Mean Reversion https://www.mql5.com/en/market/product/182363