Risk-Adjusted Returns in DeFi: Sharpe, Sortino, and Calmar for 24/7 Markets

Key takeaways

• Sharpe, Sortino, and Calmar still work. The calendar does not. Use 365 days.

• Sharpe uses total volatility. Sortino isolates downside. Calmar uses max drawdown.

• A 24/7 market produces more observations than a cash-equity book. Do not import a 252-day habit.

• Rolling windows show regime changes that a single full-history ratio hides.

The 365-Day Convention

Every risk metric in DeFi requires a decision about annualization. Traditional finance uses 252 trading days because equity markets are closed on weekends and holidays. DeFi protocols run every hour of every day, including weekends and bank holidays.

Using 252 understates annualized volatility and overstates annualized returns for DeFi portfolios by a factor of √(365/252) ≈ 1.20. A DeFi fund reporting Sharpe ratios annualized with 252 days is systematically overstating its risk-adjusted performance by approximately 20% relative to a TradFi fund computed on the same basis.

The correct DeFi convention is 365 calendar days. All risk metrics should be computed and reported on this basis to be comparable across strategies.

Sharpe Ratio

The Sharpe ratio measures excess return per unit of total volatility:

Sharpe = Mean daily return / σ_daily × √365

The risk-free rate is set to zero. There is no Treasury equivalent in 24/7 DeFi markets: crypto markets never close, and any on-chain reference rate (stablecoin lending, etc.) is itself a risk asset rather than a risk-free benchmark. Setting it to zero is both the technically correct choice and the most conservative one for LP reporting.

Volatility uses the sample standard deviation with Bessel’s correction (dividing by n − 1), applied to daily returns computed from daily NAV. The ratio requires at least two observations; it returns zero when volatility rounds to below 10⁻⁶, which flags flat returns rather than misrepresenting them as infinite Sharpe.

A Sharpe above 1.0 in DeFi is strong. Above 2.0 should be examined carefully: very high Sharpe ratios are often a sign of embedded tail risk (delta-neutral strategies with significant liquidation exposure in extreme scenarios) rather than genuine alpha.

Sortino Ratio

The Sortino ratio uses the same numerator as Sharpe but replaces total volatility with downside volatility:

Sortino = Mean daily return / σ_downside × √365

The downside denominator is the root-mean-square of all negative daily returns: computed over the total observation count n (not just the count of negative days). This is the population RMS of negative returns, not the sample standard deviation of a filtered subset. The result returns zero when no negative returns have been observed, rather than producing a division-by-zero error that would misrepresent the position.

For DeFi yield strategies, Sortino is often more representative than Sharpe. A stablecoin lending position has very low downside volatility: the position value falls only in a protocol failure scenario: but the daily rate fluctuation creates upside variance. Sharpe penalizes that upside variance; Sortino correctly ignores it.

Calmar Ratio

The Calmar ratio compares annualized return to maximum drawdown:

Calmar = IRR-APR / |Max Drawdown|

where IRR-APR is the annualized time-weighted return (TWR × 365 / days since inception). Calmar returns zero when |max drawdown| is below 0.1%: a position with no meaningful decline would produce an artificially infinite ratio. This is a guard against data noise, not a business decision.

The Calmar ratio uses since-inception metrics rather than a rolling 36-month window. DeFi positions typically have shorter lifetimes than the 36-month window that traditional fund analysis uses, so since-inception is the correct scope.

Calmar captures what Sharpe and Sortino miss: the depth of the worst historical loss. A fund with an excellent Sharpe but a 60% max drawdown scores low on Calmar: correctly signaling that the tail risk was severe even if average volatility looked acceptable.

Max drawdown must be computed on daily NAV. A fund that fell 40% intra-month and recovered before month-end would not show that drawdown in monthly reporting: only daily NAV tracking captures the actual worst-case experience.

Rolling Metrics and Regime Detection

Point-in-time risk metrics are less useful than rolling windows for understanding how a strategy’s risk profile evolves. A rolling 90-day Sharpe ratio plotted over the fund’s history shows whether the strategy consistently delivered risk-adjusted returns or whether the full-history Sharpe was driven by a single exceptional period.

For allocators evaluating a DeFi fund, the rolling Sharpe and Sortino are as important as the since-inception figures. They reveal whether the risk profile is stable or whether the manager’s edge is regime-dependent.