Calculators

Percentile Rank Calculator

Find the percentile rank of a target in a dataset with midrank, strict or inclusive tie handling, privately in your browser.

Find the percentage of dataset values below a target, with your choice of tie handling. Your data stays in your browser.

Separate 1 to 10000 finite signed numbers with newlines, commas, semicolons, or spaces. Use decimal points without thousands separators. Scientific notation is supported.

Your percentile rank will appear here.

How to calculate percentile rank

Let B be the count below the target, E the count equal to it, and N the dataset size. Midrank = 100 × (B + E/2) / N; strict rank = 100 × B / N; inclusive rank = 100 × (B + E) / N. For 2, 4, 6, 20 and target 6, these give 62.5%, 50% and 75%. A target need not occur in the data. Targets below all values return 0% and targets above all values return 100%. Ties use exact numeric equality. Enter 1 to 10000 finite signed values separated by commas, spaces, semicolons or newlines; scientific notation is supported but thousands separators are not. This is an empirical dataset rank, not a normal-distribution estimate or a calculator for the value at a given percentile. Conventions differ between applications, so choose the method required for your work. Browser floating-point results display up to 12 significant digits. All processing stays in your browser.

Frequently Asked Questions

Everything you need to know about this tool, how it works, and privacy.

How is percentile rank calculated?

Count B values below the target and E equal values out of N observations. The default midrank is 100 × (B + E/2) / N. For 2, 4, 6, 20 and target 6, the rank is 62.5%.

Why do tie methods give different results?

Strict rank counts only values below the target. Inclusive rank also counts every tie. Midrank counts half the ties. If all observations equal the target, these yield 0%, 100% and 50% respectively.

Must the target appear in the dataset?

No. A target between observations has the same result under all three methods because there are no ties. A target below all observations returns 0%; one above all observations returns 100%.

Is this the same as a z-score percentile?

No. This tool compares the target directly with your dataset without assuming a distribution. A z-score percentile estimates a position using a normal-distribution model.

What input formats and privacy are supported?

Enter up to 10000 finite signed numbers separated by commas, spaces, semicolons or newlines. Use decimal points without thousands separators. Scientific notation is accepted. Data stays in your browser and results display up to 12 significant digits.