Formula, in Risk & Portfolio Construction
Downside Deviation Formula
Downside deviation is the square root of the average squared shortfall below a target return, where the average is taken over all observations rather than only the below-target ones. It isolates harmful volatility, ignoring upside dispersion entirely, and is the denominator of the Sortino ratio.
4 variables. Published May 26, 2026. By AI Fin Hub Research.
DD = sqrt( (1/n) x sum_i ( min(0, R_i - T) )^2 ) Variables
R_i
Period return
The strategy's return in period i, expressed in the same units as the target.
T
Target return (MAR)
Minimum acceptable return. Returns at or above T contribute zero through the min(0, .) clamp. Setting T = 0 measures deviation below break-even; setting T to the risk-free rate measures deviation below the safe alternative.
min(0, R_i - T)
Clamped shortfall
The deviation below target, floored at zero. Periods that meet or beat the target add nothing, so only the magnitude and frequency of underperformance drive the result.
n
Number of observations
Total count of periods. Dividing by the full count (not just the count of shortfalls) is the defining choice that keeps downside deviation comparable across samples and makes it the proper denominator for the Sortino ratio.
Step by step
- 1
Pick the target return T that defines a shortfall.
Use T = 0 to treat any loss as downside.
- 2
For every period, compute the return minus the target and replace positive values with zero.
Returns of +2%, -1%, +3%, -5%, +1% against T = 0 give clamped shortfalls 0, -1%, 0, -5%, 0.
- 3
Square the clamped shortfalls and sum them.
Squares are 0, 0.0001, 0, 0.0025, 0; sum = 0.0026.
- 4
Divide the sum by the total number of observations n, then take the square root.
0.0026 / 5 = 0.00052; sqrt(0.00052) = 0.0228 = 2.28%.
Worked example
Five monthly returns, target T = 0
Monthly returns
+2%, -1%, +3%, -5%, +1%
Target T
0%
Observations n
5
Clamped shortfalls: 0, -0.01, 0, -0.05, 0. Squares: 0, 0.0001, 0, 0.0025, 0. Sum = 0.0026. Divide by n = 5: 0.00052. Downside deviation = sqrt(0.00052) = 0.02280 = 2.28%.
Downside deviation of about 2.28% per month. The single -5% month dominates the figure; the three positive months and the small -1% month barely register, which is exactly the asymmetry the measure is built to capture.
Common variations
Run the numbers next
- Calculators Returns Distribution Analyzer
Paste a returns CSV. Histogram, normal QQ plot, skewness, excess kurtosis, Jarque-Bera test, tail-weight index. See why Sharpe alone misleads.
- Calculators Risk-Adjusted Returns Calculator
Paste a returns CSV. Sharpe, Sortino, Calmar, Omega, alpha, beta, tracking error, information ratio, max drawdown, and tail moments — plus.
Sources
- Performance Measurement in a Downside Risk Framework , Frank A. Sortino and Lee N. Price, Journal of Investing (1994)
- Safety First and the Holding of Assets , A. D. Roy, Econometrica (1952)
Keep the topic connected
- Formula Sortino Ratio Formula
The Sortino ratio formula: excess return over a target divided by downside deviation, so only harmful volatility is penalized. With a worked example.
- Learn Sortino Ratio
Sortino ratio: same numerator as Sharpe, denominator only counts downside volatility. When it's the right number to look at.
- Formula Annualized Volatility Formula
The annualized volatility formula scales per-period return standard deviation by the square root of periods per year. Root-time scaling explained.
- Learn Sharpe vs Sortino
Sharpe vs Sortino: when the gap between the two tells you something real about a strategy's tail behaviour — and when it's just noise from a small sample.