Pythonintermediatepythonmetricsrisk

Max Drawdown in Python

A vectorized NumPy/pandas function to compute the maximum peak-to-trough drawdown of an equity curve.

1 min

Maximum drawdown is the largest peak-to-trough decline of an equity curve. This vectorized implementation runs in a single pass over a pandas Series.

max_drawdown.py

import pandas as pd
 
def max_drawdown(equity: pd.Series) -> float:
    """Return the maximum drawdown as a positive fraction (0.25 == 25%)."""
    if len(equity) < 2:
        raise ValueError("equity needs at least two points")
    running_peak = equity.cummax()
    drawdowns = (running_peak - equity) / running_peak
    return float(drawdowns.max())
 
 
if __name__ == "__main__":
    curve = pd.Series([100, 120, 90, 110, 60, 80])
    print(f"Max drawdown: {max_drawdown(curve):.1%}")  # 50.0%

How it works

cummax() gives the running peak at each point. Subtracting the equity and dividing by the peak yields the drawdown at every bar; the maximum of that series is the answer. Recovering from a drawdown d requires a gain of d / (1 - d) — a 50% drawdown needs a 100% gain to get back to even.

Try the interactive version on the Max Drawdown Calculator.