RISK · 8 MIN READ
Risk management for automated trading
A system with a genuine edge can still destroy an account through sizing alone. The reverse — a bad strategy saved by good risk management — does not happen, but the first case is common enough to be the default failure.
Risk per trade, and why it is a percentage
Expressing risk as a percentage of equity makes position size shrink automatically during a losing run and grow during a winning one. That single property prevents the most common way accounts are destroyed, which is increasing size to recover losses.
| Risk per trade | Cost of 10 consecutive losses | Gain needed to recover |
|---|---|---|
| 1% | ~9.6% | ~10.6% |
| 2% | ~18.3% | ~22.4% |
| 5% | ~40.1% | ~67.0% |
| 10% | ~65.1% | ~186% |
Ten consecutive losses is not a tail event for a system winning 40% of its trades — it should be expected within a few hundred trades. The table is the argument for the conventional 1% to 2% range.
The two things that override your setting
The broker minimum lot puts a floor under position size. If your configured percentage implies less than 0.01 lots, the platform rounds up and your real risk exceeds the setting — and it fails in the dangerous direction, on the accounts least able to absorb it.
Stops are also not exact. A stop caps the loss at roughly its level under normal conditions and only approximately during a gap or a news spike, so every risk figure is a best case rather than a guarantee.
Drawdown tolerance is a real constraint
Recovery arithmetic is asymmetric: a 20% drawdown needs a 25% gain, 50% needs 100%, and 80% needs 400%. Controlling drawdown matters more than maximising return, because the deeper the hole the less plausible the climb out.
Our own validated maximum drawdown is roughly 20% across two separate test periods. An account whose owner cannot watch it fall by a fifth without switching the system off is too small in a sense no calculator captures — and switching off at the bottom converts a temporary loss into a permanent one.
Controls that refuse rather than adjust
- A hard daily loss lockdown that halts trading for the day rather than trading through it.
- A cap on concurrent positions, enforced rather than advisory — several correlated positions is one larger bet.
- A cooldown between trades on the same symbol, so a persistent setup does not generate an entry on every scan.
- A spread guard that refuses entry when the cost of the trade is unreasonable.
Each of these is a refusal to do something, which is why they rarely appear in marketing. In our own testing the gates removed more losing trades than the entry logic found winning ones.
Keep reading
Educational information only, not financial advice. Trading leveraged products carries substantial risk of loss. Last updated 2026-08-11.