Sigma Calculator — 543+ Free Calculator Tools
Turn any value into a z-score from mean and standard deviation — plus its percentile.
The Z-Score Calculator above standardizes any value — converting a raw score into the number of standard deviations it sits from the mean — and reports the corresponding percentile. It is the standard tool for interpreting test scores, measurements, and any data drawn from a roughly normal distribution. Follow these steps.
A practical tip: the percentile assumes your data are approximately normally distributed (bell-shaped). For test scores, heights, and measurement errors this is usually a safe assumption; for heavily skewed data like incomes or house prices, treat the percentile as a rough guide rather than gospel.
A useful self-check: enter a value exactly one standard deviation above the mean — for example, x = 80 with μ = 70 and σ = 10 — and confirm the calculator returns z = 1.00 with percentile 84.13. Then enter x = 60 (one standard deviation below) and confirm z = −1.00 with percentile 15.87. These two anchor points — 84.13 and 15.87 — are worth memorizing, because they bracket the middle 68% of any normal distribution. If your own data produce wildly different percentiles for z = ±1, the inputs (not the math) are suspect.
When the population standard deviation is unknown — the usual situation in real research — substitute the sample standard deviation s computed from your data; for large samples the z-score is barely affected, and the Standard Deviation Calculator linked below finds s in seconds. One caveat: with small samples (fewer than about 30 observations), statisticians switch from the normal curve to the t-distribution for percentile lookups. The z-score formula itself never changes — only the table you read the percentile from — so the calculator’s z value remains correct regardless.
A z-score — also called a standard score — tells you how many standard deviations a value lies above or below the mean (μ) of its distribution. If the mean is the center of the data and the standard deviation (σ) is the typical distance from the center, then the z-score is simply the value’s distance from center expressed in those typical units. A z-score of 1.5 means “one and a half standard deviations above average”; a z-score of −2 means “two standard deviations below average.”
Standardization was popularized alongside the normal distribution — the bell curve described mathematically by Gauss and Laplace — because it makes wildly different measurements comparable. An SAT score of 1200 and an ACT score of 28 live on different scales, but their z-scores (1.0 and 1.4 against their respective distributions) can be compared directly. This trick, called standardization, underlies modern statistics: hypothesis tests, confidence intervals, and machine-learning feature scaling all begin by converting raw numbers to z-scores.
The essential vocabulary:
Z-scores appear wherever performance must be judged relative to a group: standardized testing (SAT, GRE, IQ), pediatric growth charts (a child’s height as a z-score against age norms), manufacturing quality control (flagging parts more than 3 sigma from spec), finance (returns measured in standard deviations), and sports analytics. Anywhere you hear “above average,” a z-score can say exactly how far above.
Two industrial-strength uses show the z-score’s range. In manufacturing, a process running at “six sigma” quality keeps critical dimensions within ±6 standard deviations of target — a defect rate of about 2 per billion opportunities, the standard Motorola made famous. In machine learning, feature standardization converts every input column to z-scores before training, so that a feature measured in thousands (say, income) does not drown out one measured in fractions (say, interest rate); gradient-based optimizers converge far faster on standardized features. Both are the same idea as Maria’s exam below: express every measurement in units of its own variability.
The z-score formula is beautifully simple: z = (x − μ) / σ. Subtract the mean to center the value at zero, then divide by the standard deviation to express the distance in spread-units. The percentile comes from Φ(z), the standard normal cumulative distribution function — the area under the bell curve to the left of z, multiplied by 100.
Worked example 1 — above average. A student scores x = 85 on an exam where the class mean is μ = 70 and the standard deviation is σ = 10. z = (85 − 70) / 10 = 15 / 10 = 1.5. The standard normal table gives Φ(1.5) = 0.9332, so the percentile is 93.32 — the student outperformed about 93% of the class. The calculator above shows exactly this: z = 1.50, percentile 93.32.
Worked example 2 — below average. Another student scores x = 60 with the same μ = 70, σ = 10. z = (60 − 70) / 10 = −10 / 10 = −1.0. By symmetry, Φ(−1) = 1 − Φ(1) = 1 − 0.8413 = 0.1587, so the percentile is 15.87. A z-score of −1.0 is the exact mirror of +1.0: 15.87% below versus 84.13% below (100 − 15.87). Notice the sign carries the entire meaning — positive is above average, negative below.
Memorize the empirical-rule landmarks: z = ±1 covers the middle 68% (percentiles 15.87 to 84.13), z = ±2 covers 95% (2.28 to 97.72), and z = ±3 covers 99.7% (0.13 to 99.87). If a z-score lands beyond ±3, double-check the inputs before believing it — genuine 3-sigma events are rare, and data-entry slips are common.
Worked example 3 — a fractional z-score. A runner logs x = 72 minutes for a race where the club mean is μ = 65 minutes and the standard deviation is σ = 8 minutes. z = (72 − 65) / 8 = 7 / 8 = 0.875. The normal table gives Φ(0.875) ≈ 0.8092, so the percentile is about 80.92 — slower than roughly 81% of the club, since a higher time is worse here. Note the sign convention still works: a lower time would give a negative z, which in racing is the desirable direction. Always check which direction is “good” before celebrating a high percentile.
Maria scores 85 on her chemistry final. The class mean is 70 with a standard deviation of 10. Her raw percentage — 85% — sounds good but says nothing about her standing; the z-score tells the real story. z = (85 − 70) / 10 = 1.5, and Φ(1.5) = 0.9332 puts her at the 93.32nd percentile: she beat roughly 93 out of every 100 classmates.
This is why teachers curve with z-scores rather than raw points. On a brutal exam where the mean is 45 and the standard deviation is 8, a score of 57 gives z = (57 − 45) / 8 = 1.5 — the identical standing as Maria’s 85, despite the raw score looking far worse. Conversely, an easy exam with mean 90 and standard deviation 4 makes a 95 into z = (95 − 90) / 4 = 1.25, only the 89.44th percentile. The z-score strips away the exam’s difficulty and reveals the performance. Enter Maria’s numbers (85, 70, 10) in the calculator above and confirm z = 1.50 with percentile 93.32.
Percentiles also set fair cutoffs. If the instructor wants the top 10% to earn an A, she needs the z-score with Φ(z) = 0.90, which is z ≈ 1.28. The cutoff score is x = μ + zσ = 70 + 1.28(10) = 82.8 — so 83 and above earns an A. For the top 25% (z ≈ 0.67), the cutoff is 70 + 6.7 = 76.7. Working backward from percentile to raw score this way — x = μ + zσ — is the inverse of the calculator’s job, and it is how every curved grading scale is actually built.
Two students apply to the same program with different tests. Ana scores 1200 on the SAT (mean 1000, standard deviation 200); Ben scores 28 on the ACT (mean 21, standard deviation 5). Raw scores cannot be compared — but z-scores can. Ana: z = (1200 − 1000) / 200 = 200 / 200 = 1.0, and Φ(1.0) = 0.8413 → 84.13rd percentile. Ben: z = (28 − 21) / 5 = 7 / 5 = 1.4, and Φ(1.4) = 0.9192 → 91.92nd percentile. Ben’s performance is stronger relative to his test’s population, even though 28 “looks” smaller than 1200.
Admissions offices, employers, and researchers make exactly this comparison whenever candidates arrive with different credentials. The requirements are modest: each score needs its own mean and standard deviation, and both distributions should be roughly normal. When those hold, the higher z-score wins — no conversion tables or guesswork needed. Try it: enter 1200, 1000, 200 for Ana and 28, 21, 5 for Ben in the calculator above, and watch the percentiles (84.13 vs. 91.92) settle the debate.
The comparison breaks down only when the distributions differ in shape. If one test’s scores are heavily skewed (say, a ceiling effect piles scores near the maximum), its percentiles compress at the top and the z-score comparison favors the normal test. In practice, major standardized tests are engineered to be approximately normal precisely so that z-score comparisons remain valid — one more reason the method is trusted in high-stakes decisions.
The five mistakes below are the ones instructors flag most often on z-score assignments. None requires advanced math to avoid — just careful reading of what each number represents.
Round out your statistical toolkit with these Sigma Calculator tools:
This page provides a free Z-Score Calculator computing z = (x − μ) / σ and the corresponding percentile via the standard normal CDF Φ, for any real x, μ, and σ > 0. Verified reference values: (85, 70, 10) → z = 1.50, percentile 93.32; (60, 70, 10) → z = −1.00, percentile 15.87; (70, 70, 10) → z = 0, percentile 50; (1200, 1000, 200) → z = 1.0, percentile 84.13; (28, 21, 5) → z = 1.4, percentile 91.92. Empirical rule: ±1σ ≈ 68%, ±2σ ≈ 95%, ±3σ ≈ 99.7%. Percentile interpretation assumes an approximately normal distribution. All calculations run locally in the visitor’s browser with no signup.
Sources: C. F. Gauss and P.-S. Laplace on the normal distribution; standard statistics texts (Moore, McCabe & Craig; Freedman, Pisani & Purves). Content reviewed for accuracy in October 2026.
Enter the value x, the distribution mean μ, and the standard deviation σ. The calculator returns the z-score and the percentile — what share of values fall below yours in a normal distribution.
Formula: z = (x − μ) ÷ σ. Example: x = 85, μ = 70, σ = 10 → z = 1.50, the 93.32 percentile — 1.5 standard deviations above the mean.
Subtract the mean from the value and divide by the standard deviation: z = (x − μ) ÷ σ. A score of 85 with mean 70 and SD 10 gives z = 1.5.
The value sits 1.5 standard deviations above the mean — about the 93rd percentile, higher than roughly 93% of values.
Yes — a negative z-score means the value is below the mean. z = −1 is one standard deviation below average, around the 16th percentile.
Comparing scores from different distributions (like two different exams), spotting outliers, and converting to percentiles and p-values.
The z-score measures distance from the mean in standard deviations; the p-value converts that distance into a probability under the normal curve.
Zero. Plugging x = μ into z = (x − μ) / σ gives 0 / σ = 0, and the 50th percentile — exactly average, by definition.
Convert each score with its own mean and standard deviation, then compare the z-scores. SAT 1200 (mean 1000, SD 200) gives z = 1.0 (84.13 percentile); ACT 28 (mean 21, SD 5) gives z = 1.4 (91.92 percentile) — the ACT performance is relatively stronger.
A common rule flags |z| > 3 as an outlier — beyond 99.7% of normally distributed data. For example, x = 105 with mean 70 and SD 10 gives z = 3.5, well into outlier territory. Some fields use |z| > 2.5; always check your data entry first.
Also try our P-Value Calculator and Standard Deviation Calculator.