PERCENTRANK in WPS Spreadsheet: Percentile Positions, Interpolation and Display Precision

• Introduced: documented by 2019-06-18 in the official WPS lesson; the first release is not established.
• Availability: the lesson identifies WPS Office for Windows Personal, without a numeric release range.
• Scope: exact current supported builds and other WPS clients remain unverified. Historical naming does not establish retirement.
• Checked: October 14, 2026. Documentation and independent calculations; no native WPS test was performed.

=PERCENTRANK(array, x, [significance])

array is the comparison dataset, x is the target, and significance controls the returned precision. The lesson documents #NUM! for an empty array or significance below 1.

The examples use distinct numeric values and explicit significance 6 unless demonstrating an exactly representable default result. Duplicate-target tie rules, targets outside the observed range and exact native precision/truncation boundaries remain unverified here.

Examples use English names and comma separators. Do not multiply an already formatted percentage again merely because its underlying value is a fraction.

Enter 10, 20, 40, 70 and 100 in A2:A6. In B2 enter and fill down:

=PERCENTRANK($A$2:$A$6,A2,6)

The independently modeled expected positions are:

For these five distinct ordered values, the inclusive rank model assigns position (rank-1)/(5-1). Four equal rank intervals connect the minimum and maximum, even though the numeric gaps between values differ.

The source documents the percentile-rank purpose and interpolation; these complete outputs are independent mathematical controls rather than captured native WPS results.

For target 70:

=PERCENTRANK(A2:A6,70,6)
=70/MAX(A2:A6)

The expected rank fraction is 0.75, while the fraction of the maximum is 0.7. They answer different questions: relative position in the ordered set versus numerical magnitude compared with 100.

Adding 10 to every observation and the target leaves their ordering and the independent rank fraction unchanged. The magnitude ratio instead becomes 80/110, approximately 0.727272727.

A percentile rank does not measure a percentage improvement, a probability of passing or a score’s distance from the mean without additional definitions.

Keep the original five observations and target 30, halfway between 20 and 40:

=PERCENTRANK(A2:A6,30,6)

The independent linear-interpolation model gives 0.375, or 37.5%. The two neighboring rank fractions are 0.25 and 0.5. Halfway between their values corresponds to halfway between those fractions:

0.25 + (30-20)/(40-20) * (0.5-0.25) = 0.375.

For target 25, one quarter of the way from 20 to 40, the same model gives 0.3125. These targets are inside the observed range; they do not certify extrapolation or a particular error outside it.

Append numeric 200 to the dataset. The ordered set now has six observations. Target 70 remains the fourth value, but the independent inclusive rank fraction becomes (4-1)/(6-1) = 0.6.

The score itself did not fall. The comparison population changed. Keep the reference range anchored and explicitly state whether each report uses a fixed benchmark population or a changing current dataset.

A shifting reference caused by an unanchored copied formula can create a similar apparent change unintentionally. Inspect the range before interpreting the rank as a substantive performance change.

For target 40 in the original dataset, =PERCENTRANK(A2:A6,40) has an exactly representable expected fraction 0.5. Displaying that cell as a percentage shows 50% according to the chosen number format.

The dated lesson’s separate walkthrough uses two-place significance and reports 77% for its own ten-score fixture. That example is not treated as a complete native rounding rule for every repeating fraction. The controls here avoid relying on an undocumented precision boundary.

Use RANK for ordinal positions and QUARTILE for selected data-value thresholds. A rank fraction and a value threshold are different outputs. Mixed reference types, singleton datasets and full current-build availability remain unverified.

• WPS Academy: PERCENTRANK, 2019-06-18 (Chinese).
• Editorial testing standards.

Worked outputs are independently calculated expectations, not captured native results. Function-specific source gaps and product/client boundaries remain explicitly scoped.