BETA.INV Function in ONLYOFFICE Spreadsheet Editor: Syntax and Examples

=BETA.INV(0.25,1,1) is expected to return 0.25. On support [2,6], =BETA.INV(0.25,1,1,2,6) is expected to return 3. The function answers “which value has this cumulative probability?” rather than “what is the probability at this value?”

The official BETA.INV reference and BETAINV reference describe the inverse beta calculation. In the version-pinned implementation, BETAINV directly inherits BETA.INV’s calculation and argument count. This guide covers both spellings; they do not need separate worked calculations.

=BETA.INV(probability, alpha, beta, [A], [B])
=BETAINV(probability, alpha, beta, [A], [B])

Use A < B. Although the help calls A and B bounds of “probability,” the implementation applies them to the returned value, using A + z(B−A). Probability remains on the 0-to-1 scale when the modeled variable is hours, a proportion or another bounded quantity. To obtain the 25th percentile, enter 0.25 or a true numeric 25% value, not 25.

The documented probability lower limit excludes zero, while the inspected source guard permits zero before invoking its numerical inverse helper. This guide uses interior probabilities and makes no native endpoint-compatibility claim. Avoid depending on p=0 or p=1 portability without testing the actual application/version.

1. Enter probabilities 0.25, 0.5 and 0.75 in A2:A4.
2. Enter alpha 1 in B2 and beta 1 in C2.
3. Enter lower endpoint 2 in D2 and upper endpoint 6 in E2.
4. In F2 enter =BETA.INV(A2,$B$2,$C$2,$D$2,$E$2) and fill through F4.
5. Label column F Threshold (hours); format A as percentages if desired.

With alpha=beta=1 the distribution is uniform, so the inverse is a straight-line transformation. A second control uses alpha=2,beta=1, whose CDF is z²: =BETA.INV(0.25,2,1) is expected to return 0.5, since the nonnegative square root of 0.25 is 0.5. Numerical inversion may produce a value differing in the last floating-point digits; compare with a suitable tolerance rather than a particular display width.

For the same uniform 2-to-6-hour model, use =BETA.INV(0.1,1,1,2,6) and =BETA.INV(0.9,1,1,2,6) for expected endpoints 2.4 and 5.6. The interval contains the central 80% of this assumed distribution, leaving 10% in each tail.

This is a model-based quantile interval, not automatically a confidence interval for an estimated mean and not proof that 80% of a small observed sample will fall inside it. Choosing or estimating alpha and beta is a separate modeling step. Retain the assumptions with the threshold values so a later user does not treat a convenient uniform example as a fitted forecast.

The source explicitly rejects probabilities below 0 or above 1, nonpositive shapes and A ≥ B with #NUM!. For example, =BETA.INV(1.1,1,1), =BETA.INV(0.5,0,1) and =BETA.INV(0.5,1,1,2,2) follow those guards. A helper-reported failure to bracket the inverse also returns #NUM!. The refinement loop has a fixed iteration limit without a corresponding failure flag, so a returned result is not a universal guarantee of numerical convergence. Valid-looking arguments do not guarantee all extreme parameter combinations will be numerically easy.

=BETAINV(0.25,1,1,2,6) has the same source-derived expected result, 3, as its modern spelling. This shared implementation does not prove identical compatibility metadata in every file format or older build. Keep the exact function identity when discussing an existing workbook and verify imports in the destination application. Neither the current help listing nor a legacy name alone establishes retirement or deprecation.

The official BETA.INV reference was checked on October 18, 2026. The implementation discussion uses official SDKJS source tagged v10.0.0.169, matching the version/build designation of Desktop Editors 10.0.0-169. Desktop Editors 10.0 was released on October 13, 2026. The first version introducing this individual function is not established by these sources. Current documentation does not mark it retired or deprecated.

Examples in this guide are expected results from documentation, source inspection and independent calculations, not results observed in the application. They have not been tested in the application or its calculation engine. The unversioned help and pinned source do not establish identical behavior in older versions, mobile clients, every server deployment or imported workbooks. Examples use English names and comma argument separators; adapt formula-entry separators to your editor settings.

• AVERAGE