Risk management for algo traders

Most new algo traders begin by asking whether a strategy wins. The more useful first question is how much the strategy can lose while it is doing exactly what the backtest says it will do.

A losing streak is not automatically evidence that an algorithm is broken. It is part of the distribution of outcomes. Risk management gives that distribution room to play out without one bad session deciding the fate of the account.

Daily drawdown measures how much the account has lost during a trading day. Overall or maximum drawdown measures the decline from a previous high-water mark. They are different limits and a system can be inside one while being close to breaching the other.

Prop-firm traders should read the current rules directly from the firm, including whether floating profit and loss, commissions, swaps, and trailing limits count. Never rely on a remembered percentage from an older rule set.

The same entry logic can behave conservatively or aggressively depending on position size. A simple risk-per-trade calculation starts with the cash amount you are prepared to lose, then divides it by the stop distance and the instrument's value per point.

Because broker contract specifications differ, verify tick size, tick value, minimum volume, volume step, and margin requirements inside your own MT5 account. A setting that is safe on one symbol suffix may be wrong on another.

An automated system still needs an operator. Review the account balance and equity, open exposure, execution logs, connection status, and whether the terminal is receiving fresh prices. A quiet bot can be normal; a quiet bot with no heartbeat or stale prices is an operational problem.

Keep a record of configuration changes. If you change risk, symbol, broker, or time settings, mark the date so later results are not mixed together as though they came from one system.