The equity curve plots account balance through the backtest window. The shape carries information the headline metrics don't.
Shapes to look for
Monotone up
Smooth, consistent grind. Healthy. Pair with high trade count and the strategy is plausibly real.
Lumpy / step-function
Long flat stretches punctuated by a few big winners. Often overfit - the strategy works only in specific conditions.
Late pump
Flat for most of the window, hockey-stick at the end. Could be a genuine regime shift the strategy caught, or curve-fitting to recent data. Re-run with a date range ending earlier to test.
Front-loaded
Big move early, flat or declining since. Common after a one-time discovery. Walk-forward analysis exposes this.
Drawdown craters
Sudden vertical drops. Identify the trades and look for stop failures, gap events, or correlation with news.
Together with drawdown
The deepest drawdown on the equity curve = the worst run the strategy had. Ask yourself: would I have kept running through it? If no, the strategy isn't deployable even if expected return is positive.
