Should I stop a strategy that's in a drawdown?

First separate 'drawdown within design' from 'behavior outside design.' Every strategy's backtest shows its historical worst losses — if the live drawdown is inside that envelope and trades still match the rules, stopping now is selling the strategy at its low, the exact emotional error automation exists to prevent. Stop immediately, though, when behavior diverges from design: trades the rules can't explain, frequency wildly off backtest, or a market structurally outside anything tested (its ticker's fund closing, circuit-breaker regimes). A useful pre-commitment: write down, before deploying, the drawdown at which you'll stop regardless — say 1.5x the backtest's max — and let that number, not a bad week, make the call.