PlanMaker PERCENTILE.EXC: Exclusive Percentiles and Interior Calculation Checks
Compare exclusive and inclusive percentile examples in PlanMaker, reconstruct interior positions and understand the limits of small-sample tail claims.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
PERCENTILE.EXC returns an exclusive percentile of a numeric dataset. Its percentage input must lie strictly between zero and one. It can differ from PERCENTILE.INC even when both formulas use the same data and the same percentage.
Availability and version history
- Introduced: Documented by the PlanMaker 2018 manual; the original introduction release and date are not established.
- Availability: Documented in PlanMaker 2026. This checked version is not a final supported-version endpoint.
- Scope: Current reference belongs to the PlanMaker 2026/NX manual family. FreeOffice and individual mobile-build availability are not established here.
- Checked: October 14, 2026, against the official PERCENTILE.EXC reference and historical manual. Mathematical reconstructions are explicitly scoped; no native PlanMaker test was performed.
Syntax and the documented contrast
=PERCENTILE.EXC(Range, k)
Both arguments are required. The official reference specifies numeric data and 0 < k < 1, excluding zero and one. It identifies k=0.5 as the median and distinguishes this function from the inclusive PERCENTILE/PERCENTILE.INC pair.
On VendorExclusive, enter 1 through 9 in A1:A9. In C1 enter =PERCENTILE.EXC(A1:A9,0.75) for the documented result 7.5. In C2 enter =PERCENTILE.INC(A1:A9,0.75) for the separately documented 7. This difference is deliberate; changing the name changes the percentile convention.
The interior interpolation model
The checks below use an independent exclusive model with position h=(n+1)k within sorted observations. At an interior position j+d, interpolate from position j toward j+1 by fraction d. For the vendor example, n=9 and k=0.75 give h=7.5, halfway between seven and eight.
The model matches that published example, but the short PlanMaker page does not state a full algorithm or extreme-tail rules. NIST’s percentile discussion describes this interior interpolation convention. Its own endpoint treatment is not evidence of PlanMaker’s behavior, so the examples stay inside the observed order-statistic positions.
Compare three useful interior percentiles
Create ExclusiveTimes using A1:H8. Put Sorted hours in A1 and enter 12,18,25,35,40,60 in A2:A7. Enter headers k, Exclusive percentile, Inclusive comparison, Exclusive position, Inclusive position and Difference in C1:H1. Put 0.25,0.5,0.75 in C2:C4.
- D2:
=PERCENTILE.EXC($A$2:$A$7,C2) - E2:
=PERCENTILE.INC($A$2:$A$7,C2) - F2:
=(COUNT($A$2:$A$7)+1)*C2 - G2:
=1+(COUNT($A$2:$A$7)-1)*C2 - H2:
=D2-E2
Fill D2:H2 through row 4.
| k | Exclusive position | Inclusive position | Expected D under the model | Expected E under the model | Expected H |
|---|---|---|---|---|---|
| 0.25 | 1.75 | 2.25 | 16.5 | 19.75 | -3.25 |
| 0.5 | 3.5 | 3.5 | 30 | 30 | 0 |
| 0.75 | 5.25 | 4.75 | 45 | 38.75 | 6.25 |
The median agrees because both position formulas give 3.5. The lower and upper quartile positions differ, so the corresponding values differ too. In C6 enter =MEDIAN(A2:A7) for the separate control 30.
Reconstruct the upper percentile
Use J1:K9 for labels and values on ExclusiveTimes. Put k=0.75 in K2, then K3 =COUNT(A2:A7) → 6; K4 =(K3+1)*K2 → 5.25; K5 =INT(K4) → 5; K6 =K4-K5 → 0.25.
In K7 enter =INDEX(A2:A7,K5,1) → 40 and K8 =INDEX(A2:A7,K5+1,1) → 60. K9 =K7+K6*(K8-K7) gives 45, a quarter of the way from 40 to 60. The output need not appear among the original observations.
Keep extreme tails separate from interior checks
For six observations, the model’s first and last observed positions correspond to k=1/7 and k=6/7. A request at k=0.9 is inside the documented 0-to-1 interval but gives h=6.3, beyond the last observed position.
The checked PlanMaker entry does not establish whether every such small-sample tail request errors, clamps or behaves in another way. This guide therefore gives no predicted native result or error token for that case. If the minimum or maximum is what you need, use MIN or MAX directly.
Change one central observation
Copy ExclusiveTimes to ChangedMiddle and change A4 from 25 to 18, leaving the sorted order intact. The data are now 12,18,18,35,40,60.
Under the stated model, D2:D4 becomes 16.5,26.5,45, E2:E4 becomes 18,26.5,38.75, and H2:H4 becomes -1.5,0,6.25. C6 becomes 26.5. The reconstruction in K still returns 45 because the two upper observations are unchanged.
This comparison shows why the positions surrounding the requested percentile matter. It is not a claim that changing any middle value can never alter an upper percentile; other sample sizes or k values can select a different pair.
Practical cautions
Use explicit numeric inputs and the same measurement units. Keep k as a fraction, and do not substitute exclusive and inclusive names without documenting the convention. Mixed types, empty-only ranges, exact error tokens and native precision remain unverified.
The Excel 2010 note in SoftMaker’s manual concerns file compatibility with that other product; it does not date PlanMaker’s introduction. Related guides: PERCENTILE.INC, PERCENTILE and MEDIAN.