Risk & Measurement
Measuring what matters before you invest
A strategy that returned 15% annually in backtesting delivered just 2.3% once it went live — and a 2014 study by Bailey and Lopez de Prado showed this kind of gap is not the exception but the norm. The culprit is not bad luck. It is bad measurement.
The metrics we trust are often misleading
The Sharpe ratio, the single most cited number in quantitative finance, assumes that returns follow a neat bell curve. In practice, they rarely do. Strategies with options-like payoffs, tail risk exposure, or leverage can produce Sharpe ratios that look outstanding right up until a catastrophic drawdown reveals what the metric was hiding. A strategy that clips small gains while absorbing rare, devastating losses can appear far safer than it actually is.
Even when the Sharpe ratio is measured correctly, selection bias inflates it. If you test a hundred strategy variations and report the best one, the winning Sharpe ratio is almost certainly overstated. The Deflated Sharpe Ratio, introduced by Bailey and Lopez de Prado, accounts for this by adjusting for the number of trials — a correction that turns many "impressive" results into statistical noise.
Backtesting is not evidence
The gap between backtested and live performance is one of the most well-documented phenomena in quantitative finance. Overfitting to historical data, ignoring transaction costs, and failing to model market impact all conspire to make paper returns look better than reality. Harvey, Liu, and Zhu argued that the traditional t-statistic threshold of 2.0 is far too low given the sheer number of hypotheses the field has tested. Their proposed threshold of 3.0 filters out the majority of false discoveries.
Measurement as foundation
Honest risk measurement is not a peripheral concern — it is the foundation on which every other decision in quantitative investing rests. Position sizing depends on volatility estimates. Strategy selection depends on drawdown analysis. Portfolio construction depends on correlation measurement. When any of these inputs is wrong, the downstream decisions compound the error.
The articles in this section dissect the tools and traps of risk measurement: what the Sharpe ratio actually tells you (and what it does not), how maximum drawdown reveals risks that variance-based metrics miss, why volatility targeting can stabilize returns, and how transaction costs silently erode backtested alpha. The goal is not to discourage quantitative analysis but to make it more rigorous — because in this field, the quality of your measurement determines the quality of your results.
Key Research Insights
The Sharpe ratio remains the most widely used risk-adjusted performance measure, but it assumes normally distributed returns — an assumption violated by most real-world strategies.
When multiple strategies are tested on the same dataset, the best-performing Sharpe ratio is almost certainly inflated. The Deflated Sharpe Ratio corrects for this selection bias by accounting for the number of trials run.
Given the hundreds of factors tested across the literature, a t-statistic of 2.0 is no longer sufficient evidence. A threshold of t > 3.0 is needed to account for multiple testing and reduce false discoveries.
Every trade moves the market. Optimal execution frameworks show that market impact costs grow with the square root of trade size, making large positions disproportionately expensive to enter and exit.
Glossary
Risk
Backtesting Pitfalls: Why Most Backtests Lie
Most backtests are too good to be true. Survivorship bias, look-ahead bias, and data snooping inflate performance, while unrealistic assumptions about costs and liquidity mask fatal flaws. Learn how to build honest backtests using deflated Sharpe ratios and walk-forward analysis.
Transaction Costs and Slippage: The Hidden Drag on Quant Strategies
Transaction costs are the single largest reason why theoretically profitable quant strategies underperform in practice. Understanding the components of execution costs — commissions, bid-ask spreads, and market impact — and applying models like Almgren-Chriss for optimal execution is essential for any serious quantitative investor.
Volatility Targeting: Scaling Risk for Better Returns
Volatility-managed portfolios scale exposure inversely to recent realized volatility. This simple approach improves Sharpe ratios across equities, bonds, and currencies without requiring any ability to forecast returns.
Maximum Drawdown: The Risk Metric Investors Fear Most
Volatility tells you about typical fluctuations, but maximum drawdown tells you about the worst pain. MDD captures the largest peak-to-trough decline in portfolio value -- the number that keeps allocators awake at night. Understanding drawdown metrics like the Calmar ratio and Conditional Drawdown at Risk is essential for realistic strategy evaluation.
The Sharpe Ratio: Measuring Risk-Adjusted Returns
The Sharpe ratio is the most widely used measure of risk-adjusted performance in finance, yet it is frequently misunderstood and misapplied. We explain its construction, assumptions, limitations, and alternatives.