Contact

FINV Function (OpenOffice Calc)

Statistical Intermediate OpenOffice Calc

Find right-tail F cutoffs with Apache OpenOffice Calc FINV, including endpoint rules, ordered degrees of freedom and variance-ratio confidence bounds.

Verification status: Official release source and independently calculated examples for FINV were reviewed. No new Apache OpenOffice Calc application execution is claimed for this guide; the stated version and example limits still apply. How formula examples are checked

Quick answer: =FINV(0.05;5;5) returns approximately 5.050329057633, the F cutoff with 5% of the reference distribution above it.

Availability and version history

  • Introduced: First release not established. The OpenOffice.org 3.3 function appendix documents this name; that is presence evidence, not its introduction date.
  • Supported in: Apache OpenOffice 4.1.16, with no final supported release established.
  • Scope: Desktop Calc; English names and semicolon separators are used below. The numeric contract refers to the cited release.
  • Checked: October 14, 2026. Retained official documentation/source and independent mathematics checked; no new Apache OpenOffice Calc execution is claimed.

Inverse right-tail probability

FINV returns the x for which FDIST(x;df1;df2)=p. Small positive p gives a large cutoff. It is an inverse-tail operation, not the reciprocal 1/FDIST and not a left-tail percentile. Function reference.

The order of the degree counts still matters. The numerator model uses DegreesFreedom1, and the denominator model uses DegreesFreedom2. Keep those labels with the input cells rather than using two unnamed numbers.

Syntax and accepted range

=FINV(Number; DegreesFreedom1; DegreesFreedom2)

All arguments are required. Number is a right-tail probability satisfying 0<p≤1. Each degree count is floored and must then satisfy 1≤df<10,000,000,000. The release implementation rejects p=0 because no finite positive-tail cutoff exists there.

At p=1 the cutoff is 0. The inverse is obtained numerically, so a permitted numeric input is not a promise of convergence or uniform extreme-tail precision. A similarly named inverse in another program may use a different tail or parameter convention.

Reproducible reference cases

Values below were independently calculated and rounded to twelve decimal places. Retain unrounded results in subsequent worksheet calculations.

Formula Expected result, approximately
=FINV(.5;5;10) 0.931933160851
=FINV(.05;5;5) 5.050329057633
=FINV(.1;5;5) 3.452982248038
=FINV(.5;5;5) 1
=FINV(.2;2;4) 2.472135955
=FINV(.05;2;4) 6.944271909999
=FINV(.95;2;4) 0.05195670417
=FINV(1;5;5) 0

Practical worksheet: determine and verify a cutoff

Put 0.05 in A2 and 5 in B2:C2. In D2 enter =FINV(A2;B2;C2), giving approximately 5.050329057633. In E2 enter =FDIST(D2;B2;C2); the expected round trip is approximately 0.05.

A value of 0.95 requests a low cutoff with most of the distribution above it. To obtain the upper 5% cutoff, enter 0.05, not 0.95. A percentage-formatted probability cell should contain 5% or 0.05, not the number 5.

Advanced example: interval for a variance ratio

For independent normal samples, let R=s1²/s2² and let d1=n1−1 and d2=n2−1. A central 95% confidence interval for the population variance ratio is:

Lower = R / FINV(0.025;d1;d2)
Upper = R / FINV(0.975;d1;d2)

The larger quantile is in the denominator of the lower bound. For R=2 and d1=d2=5, the two cutoffs are approximately 7.146381828733 and 0.139930950230; the variance-ratio bounds are approximately 0.279861900460 and 14.292763657466. This wide interval reflects the uncertainty of small samples under the stated model.

Do not reverse the samples after observing which variance is larger. If the scientific question was one-sided, choose that direction in advance rather than borrowing a two-sided interval’s tail probabilities without explanation.

Reciprocal check and errors

For interior p, FINV(p;d1;d2)=1/FINV(1-p;d2;d1). For example, =FINV(0.2;2;4) and =1/FINV(0.8;4;2) both give approximately 2.472135955000. Reverse both the probability and the degree ordering in this check.

Probabilities at or below zero, above one, or invalid degree counts follow the invalid-argument branch. Numerical inversion can separately report non-convergence. Avoid replacing either condition with an arbitrary large cutoff, and do not present these source-derived rules as freshly observed application errors.

Official evidence and calculation scope

The examples were recomputed independently against the retained release contract. Native import, locale and extreme-input behavior is not newly certified.