In investing, intuition often tells us that a high win rate must equal success. If a strategy is right nine times out of ten, surely it’s a winner. But in systematic investing, this is one of the most deceptive illusions. A strategy can win 90% of the time and still lose money overall – sometimes spectacularly.

This paradox highlights one of the most important lessons in quantitative finance: it’s not how often you win, but how much you win when you’re right, and how much you lose when you’re wrong.

The Fragile Math of Win Rates

Let’s start with a question:

“With a 90% win rate, how many losing trades could you face in a row in the worst-case scenario?”

The intuitive guess might be “not many.” The uncomfortable truth? The answer is infinity. Even with a 90% win rate, nothing prevents markets from delivering an unbroken streak of losses. That is why risk-of-ruin thinking matters more than win rates.

Consider a simple example: a strategy that makes $1 on 9 trades in a row, then loses $20 on the 10th. The win rate is 90%, but the net result is –$11. Many short-volatility, carry, or arbitrage strategies fall into this trap. They provide a stream of small, steady gains until a rare but massive loss erases years of progress.

This is the danger of asymmetric payoff structures. Strategies optimized for high win rates often expose themselves to “fat tail” risks – low-probability but high-impact losses. Investors, attracted by the comfort of frequent wins, underestimate the risk of ruin.

The opposite is also true: a strategy with only 30–40% winning trades can be highly profitable if its average gains vastly outweigh its losses. This is why systematic investors often pay less attention to win ratios and more to payoff distributions, tail risks, and risk-adjusted returns.

Research Spotlight

Classic and modern research illustrates why win rate is a misleading metric:

• • Taleb (2007, The Black Swan): Popularized the concept of rare but devastating losses that dominate long-term outcomes, particularly in strategies with high win rates.
  • Nassim Taleb & Espen Haug (2011): Showed how option-selling strategies systematically underprice tail risks, leading to attractive win rates but explosive downside.
  • Bailey & López de Prado (2012): Demonstrated why performance metrics like Sharpe Ratio and Maximum Drawdown are far more reliable indicators of strategy health than raw win percentage.
  • LSE (2022): A study of thousands of retail trading accounts found that over 75% of traders had win rates above 50%, yet most still lost money because their losses were larger than their gains.
  • RiskMetrics Research: Confirmed that selling volatility through options generated “pseudo-consistent” returns with high win rates – until sudden volatility spikes led to catastrophic losses.
    The evidence is overwhelming: win rate alone is irrelevant without context.

Omphalos Perspective

At Omphalos Fund, we design agents to think in terms of distribution, not frequency. Each trading agent is evaluated not only on how often it is correct, but on the shape of its return profile.

Key principles we apply include:

• • Payoff asymmetry: ensuring the ratio of average gain to average loss remains favorable.
  • Drawdown resilience: testing how agents behave under rare but extreme stress conditions.
  • Uncorrelation in losses: combining agents whose drawdowns don’t overlap, especially in negative market periods.

This approach means that even if individual agents show low win rates, their contribution to the total portfolio can still be strongly positive. The system isn’t about being right most of the time – it’s about surviving and compounding over time.

A Case in Point: The “Short Volatility Trap”

From 2012 to early 2018, many investors embraced strategies that sold volatility, betting on the persistence of calm markets. These strategies showed win rates well above 80%. Every day, they earned small, steady profits – until “Volmageddon” in February 2018, when volatility spiked and billions were wiped out in hours.

The lesson is stark: frequency of success is meaningless without magnitude of risk. A 90% winning record is not an edge if the 10% losses are fatal.

👉 In the next chapter, we’ll tackle another dangerous illusion: the “lie of normal distribution.” While many financial models assume bell curves and stable probabilities, markets are dominated by fat tails and outliers – the very risks most backtests and naive statistics fail to capture.