What is standard deviation?

Standard deviation (SD) quantifies how much individual values deviate from the mean on average. A small SD means values cluster tightly around the mean (consistent, predictable). A large SD means values are widely scattered (variable). Population SD formula: sigma = sqrt(sum of (x - mean)^2 / N). The squaring step ensures all deviations are positive and penalises larger deviations more heavily. Taking the square root returns the result to the same unit as the original data. Example: test scores [70, 75, 80, 85, 90] have mean 80 and SD approximately 7.07, meaning most scores are within 7 points of 80.

What is the difference between population and sample standard deviation?

Population SD uses N in the denominator and is correct when you have data on every member of the group - such as the exact ages of all 30 students in one specific classroom. Sample SD uses N-1 (Bessel's correction) and is correct when your dataset is a sample drawn from a larger population - such as measuring heights of 50 students to estimate variability across a school of 1,000. Using N instead of N-1 on a sample systematically underestimates true population variability. Most real-world analysis uses sample SD. When in doubt, use sample SD - it is slightly conservative and appropriate for any subset of a larger group.

Variance = Standard Deviation squared. For a dataset with SD of 7.07, the variance is 50. Variance is mathematically convenient because variances of independent variables add directly: Var(A + B) = Var(A) + Var(B). This additive property is why variance is used internally in statistical modelling, ANOVA, and portfolio theory. However, variance is in squared units - if your data is in kilograms, variance is in kg^2. This makes variance hard to interpret in the original measurement context. Standard deviation solves this by taking the square root, returning the result to the original unit and making it directly comparable to the data values.

What is the coefficient of variation?

CV = (Standard Deviation / Mean) x 100%. It expresses variability as a percentage of the mean, enabling comparison across datasets with different units or scales. A manufacturing process producing bolts with mean diameter 10 mm and SD 0.2 mm has CV = 2% (very consistent). Another process with mean 100 mm and SD 5 mm has CV = 5% (less consistent in relative terms, despite larger absolute SD). CV is also called relative standard deviation (RSD) in chemistry. In finance, CV = SD / mean return allows comparing volatility-per-unit-of-return across funds with different average returns. Lower CV means better risk-adjusted consistency.

How is standard deviation used in finance and investing?

Standard deviation is the primary measure of investment risk (volatility). A Nifty 50 fund with 12% average return and 18% standard deviation means roughly 68% of years had returns between -6% and +30% (mean plus or minus 1 SD), and 95% of years between -24% and +48% (mean plus or minus 2 SD). High standard deviation means high risk. Sharpe ratio = (Return - Risk-free rate) / Standard deviation - measures return per unit of risk; higher is better. In manufacturing (Six Sigma), quality control aims for plus or minus 6 standard deviations from the target - fewer than 3.4 defects per million units.

What is the 68-95-99.7 rule?

The 68-95-99.7 rule (also called the empirical rule) describes the percentage of data falling within 1, 2, or 3 standard deviations of the mean in a normal distribution. Approximately 68% of values fall within 1 SD of the mean. About 95% fall within 2 SDs. About 99.7% fall within 3 SDs. Example: a fund has mean return 12% and SD 8%. The rule predicts 68% of years produce returns between 4% and 20% (12 plus or minus 8). A return below -4% (more than 2 SDs below mean) should occur only about 2.5% of the time. Data points beyond 3 SDs are statistical outliers worth investigating. This rule applies only to approximately normal (bell-curve) distributions.

What is the Sharpe ratio and how does standard deviation factor in?

The Sharpe ratio measures risk-adjusted return: Sharpe = (Portfolio return - Risk-free rate) / Standard deviation. It tells you how much excess return you earn per unit of risk taken. Example: Fund A returns 15% with SD 12%. Fund B returns 12% with SD 6%. Risk-free rate (bank FD) = 7%. Sharpe A = (15 - 7) / 12 = 0.67. Sharpe B = (12 - 7) / 6 = 0.83. Fund B has a higher Sharpe - it delivers more return per unit of risk even though Fund A has higher absolute returns. A Sharpe above 1.0 is generally considered good; above 2.0 is excellent. Most Indian equity mutual funds have Sharpe ratios between 0.4 and 1.2 over 3-year periods.

📉

Standard Deviation Calculator

Calculate mean, variance, standard deviation, and coefficient of variation for any dataset. Population and sample modes.

About the Standard Deviation Calculator

Standard deviation measures how spread out your data is around the mean. A low standard deviation means values cluster tightly around the average; a high one means they are widely scattered. In finance, standard deviation measures portfolio volatility - a fund with 12% average return and 8% std dev is less predictable than one with 12% return and 3% std dev. In quality control, it determines whether a manufacturing process is consistent.

Standard Deviation Formula

Population SD: σ = sqrt(Σ(x - μ)² / N) · Sample SD: s = sqrt(Σ(x - x̄)² / (N-1))

μ or x̄ = mean · N = total count · Use sample SD (N-1) when working with a sample, not the full population · Variance = SD² · 68-95-99.7 rule: 68% of data falls within 1 SD of mean, 95% within 2 SD, 99.7% within 3 SD

Worked Example

Monthly returns of a mutual fund over 6 months: 2%, 5%, -1%, 8%, 3%, 4%

Returns:2, 5, -1, 8, 3, 4

Mean:(2+5-1+8+3+4)/6 = 3.5%

Deviations squared: 2.25, 2.25, 20.25, 20.25, 0.25, 0.25 · Variance = 45.5/5 = 9.1 · Sample SD ≈ 3.02% · 68% of months should fall between 0.48% and 6.52%

Tips & Insights

1
Use sample standard deviation (N-1 denominator) for any dataset that is a sample from a larger population. If you measured the heights of 50 students from a school of 1,000, use N-1. Only use population SD when you have data on every single member of the group.
2
Coefficient of variation (CV) = SD / Mean expresses variability relative to the average. CV is useful for comparing two datasets with different units or scales: a manufacturing process with CV of 2% is far more consistent than one with CV of 15%, regardless of the absolute values.
3
In finance, annualized volatility = monthly SD × sqrt(12). A mutual fund with monthly SD of 3% has annual volatility of approximately 10.4%. This is how Sharpe ratio and Value-at-Risk are computed from monthly return data.
4
The 68-95-99.7 rule: in a normal distribution, 68% of values fall within 1 SD of the mean, 95% within 2 SDs, 99.7% within 3 SDs. A data point beyond 3 SDs is statistically very rare and worth investigating as an outlier or data entry error.
5
Standard deviation is sensitive to outliers - one extreme value can dramatically inflate it. If your SD seems unusually large, check whether a single data point is far from the others. For outlier-resistant spread, use interquartile range (IQR = Q3 - Q1) instead.
6
When comparing two investment options with the same expected return, always prefer the one with lower standard deviation - it gives you the same return for less risk. The Sharpe ratio formalises this: Sharpe = (Return - Risk-free rate) / SD. A higher Sharpe means better risk-adjusted performance.
7
For quality control in manufacturing (Six Sigma), the goal is to keep the process mean at least 6 standard deviations from the nearest specification limit. At 6-sigma quality, defects are fewer than 3.4 per million opportunities - an SD-based target that transformed manufacturing quality in the 1980s and 1990s.

Why this matters for you

Standard deviation is the fundamental measure of risk in finance, quality control, and scientific measurement. When a mutual fund advertises 15% CAGR, the standard deviation tells you how much that figure varies year to year. Two funds with identical returns but different standard deviations offer very different investor experiences - and the risk-adjusted comparison (Sharpe ratio) uses standard deviation as its risk measure.

In data science and machine learning, understanding the spread of your features is a prerequisite to almost every algorithm. Normalisation (z-score = (x - mean) / SD) puts features on the same scale so that no single variable dominates the model purely due to its magnitude. Features with very high SD may need log transformation; features with SD near zero are often uninformative and can be dropped. SD is the first statistic a data scientist computes when exploring a new dataset.

Students appearing for CAT, GMAT, or statistics papers in B.Com and MBA programmes regularly encounter standard deviation problems. The conceptual distinction between sample and population SD, the relationship between variance and SD, and the 68-95-99.7 rule for normal distributions are tested explicitly. Beyond exams, any professional working with data, running experiments, or evaluating performance metrics will use SD as their go-to measure of variability - making it one of the highest-value statistical concepts to master.