Risk Management

Understanding Drawdowns: The Reality of Quantitative Investing

By Senior Portfolio Manager December 5, 2023

If a fund manager tells you they never lose money, run. In quantitative trading, drawdowns are not a bug—they are a feature. They are the cost of doing business in the pursuit of Alpha. The key is distinguishing between a normal statistical drawdown and a broken strategy.

What is a Drawdown?

A drawdown is simply the peak-to-trough decline during a specific period for an investment. It measures the "pain" an investor would have felt if they invested at the absolute top and sold at the absolute bottom.

The Mathematical Reality: To recover from a 20% loss, you need a 25% gain. To recover from a 50% loss, you need a 100% gain. This asymmetry is why capital preservation is our primary directive.

The Role of Variance

Even a strategy with a Sharpe Ratio of 2.0 (which is excellent) will have losing months. This is due to statistical variance. Market regimes change—volatility expands and contracts. A trend-following system will bleed during a choppy sideways market. This is expected.

Metric to Watch: The MAR Ratio

At Virexan Capital, we prioritize the MAR Ratio (CAGR / Max Drawdown).

  • If a strategy makes 30% a year but has a 60% drawdown, it is uninvestable (MAR = 0.5).
  • If a strategy makes 15% a year but has a 5% drawdown, it is elite (MAR = 3.0).

How We Manage Drawdowns

  1. Diversification: We trade uncorrelated assets. When Nifty tech stocks are falling, Gold might be rising.
  2. Sizing based on Volatility: When market volatility spikes (measured by India VIX), our position sizes automatically reduce to keep risk constant.
  3. The "Uncle Point": Every strategy has a pre-defined point where we turn it off. If a strategy exceeds its modeled max drawdown by 1.5x, we kill it. No questions asked.

Investing in algos is not about avoiding losses; it is about ensuring those losses are calculated, contained, and compensated for by larger wins.

Share this analysis

LinkedIn Twitter/X Email

Related Research

Explore Our Strategies

Deep dive into our core algorithmic frameworks.

Validation Methodology

How we stress-test every line of code.