GROWTH in ONLYOFFICE Spreadsheet Editor

• Introduced: Original introduction release is not established by the checked sources.
• Available: Listed in current unversioned ONLYOFFICE help, checked October 18, 2026.
• Implementation inspected: Official SDKJS v10.0.0.169, commit 3679e0375a758d7dc59d31337a45461be3525ed8. The snapshot does not prove identical behavior in every installed editor.
• Retired or deprecated: These sources do not establish retirement or formal deprecation.
• Testing: Numerical examples are independent mathematical expectations, and implementation discrepancies are labeled source predictions. No ONLYOFFICE application or calculation-engine execution was performed.

The examples use English function names and commas. Use your locale’s argument separator if different. Retain full precision in formulas; round only the displayed result. Nested output grids describe source-derived array contents. Native spill placement, obstruction errors, array-entry shortcuts and recalculation behavior have not been tested in the application. Reserve a separate output area and follow the array-formula entry method supported by your installed editor; do not assume a particular keyboard shortcut or automatic spill behavior.

=GROWTH(known_y, [known_x], [new_x], [const])

One to four arguments are registered. The one-predictor model is y=b*m^x. The inspected calculation takes natural logarithms of positive known_y values, fits a linear least-squares relationship LN(y)=LN(b)+x*LN(m), and exponentiates the fitted log responses.

TRUE or omitted const estimates b. FALSE sets b=1, because the log-scale intercept is forced to zero. It does not set the original response to zero. A multiplier m=2 means doubling per one-unit increase in x; it is not an additive increase of 2, and it is not itself a 2% rate. A multiplier of 1.02 would represent a 2% multiplicative increase per x unit.

Omitted known_x generates 1,2,…,N. Omitted new_x uses the known x positions. Use explicit x when spacing, units or the time origin matters. LOGEST returns the fitted multipliers and b, while GROWTH returns fitted or predicted y values.

1. Put new x values 3 and 4 in D1:D2.
2. Evaluate =GROWTH(B1:B3,A1:A3,D1:D2,TRUE) in an independent output area.
3. The mathematically expected vertical result is 24; 48.
4. Check directly: 3×2³=24 and 3×2⁴=48.

The log responses are LN(3), LN(3)+LN(2), LN(3)+2LN(2), so the fitted log slope is LN(2) and the intercept is LN(3). This exact pattern makes the coefficients independently checkable. Actual floating-point calculations may differ in their last digits.

=GROWTH(B1:B3) predicts the training values 3,6,12, but with implied x=1,2,3 the underlying b is 1.5 rather than 3. The same fitted training responses do not establish identical parameter interpretation. If you forecast new x values, use the same origin and spacing used to fit the model.

Using FALSE on the original x=0,1,2 data imposes y=1 at x=0, even though the observation there is 3. The fitted log slope becomes LN(2)+(3/5)*LN(3). Its multiplier is approximately 3.866364089864.

Consequently =GROWTH(B1:B3,A1:A3,D1:D2,FALSE) has independent expectations approximately 57.797392446737 and 223.465762643812. This large difference is a result of refitting under a constraint, not a formatting effect. Do not force b=1 merely to simplify the equation when the constraint is unsupported.

For a multiple-predictor example on a fresh sheet, enter predictor pairs (0,0), (1,0), (0,1), (1,1) in A1:B4 and positive responses 2,6,10,30 in C1:C4. These satisfy y=2*3^x1*5^x2. With new predictors 2,3 in E1:F1, =GROWTH(C1:C4,A1:B4,E1:F1,TRUE) is independently expected to give 2250, since 2×3²×5³=2250.

With a response column, each predictor occupies a known-x column and new_x must have the same number of predictor columns. With a response row, predictors run in rows. For one predictor, predictions follow the supplied new_x grid shape. Native placement or array-entry shortcuts are not inferred from these mathematical dimensions.

All known_y values must be strictly positive. Zero or negative responses predict internal not_numeric, conventionally #NUM!, because the model requires their logarithms. Do not replace zero with an arbitrary tiny number without recognizing that it changes the fitted problem.

Known-x/known-y shape incompatibility is checked before element types and predicts bad_reference (#REF!). Nonnumeric training elements in compatible matrices predict wrong_value_type (#VALUE!). New-x values must be numeric as well; blanks and text are not treated as automatically removable records. Multi-predictor new-x shape mismatch or failed matrix calculations can produce #VALUE! through the wrapper.

For a single predictor with identical known x values, the helper substitutes 1 for a zero centered x-square denominator. With the constant enabled, the result is a constant equal to the geometric mean of positive y values. That fallback does not identify an exponential rate. Dependent multiple predictors can fail or be numerically unstable; there is no general guarantee that redundancy will be removed correctly.

The least-squares criterion is in log y, not original y. The fit minimizes squared log residuals. Exponentiating a fitted log value is not automatically an unbiased estimate of the arithmetic mean response when errors have a multiplicative structure. The function applies no general bias correction.

An exponential trend can amplify small coefficient changes into large extrapolation differences. Keep units and origins explicit, inspect log residuals, and distinguish a descriptive curve from a justified forecast. GROWTH supplies no uncertainty interval or guarantee that future values follow the same multiplier.

Recalculate representative formulas when moving a workbook between editors. Matching names do not guarantee identical argument coercion, array behavior, numerical precision or error handling. These source-derived examples do not establish automatic spill behavior or a general import/export guarantee.

• LN
• EXP
• GEOMEAN

• ONLYOFFICE GROWTH help, checked October 18, 2026.
• Official SDKJS implementation, 2248–2615, tag v10.0.0.169 and exact commit above.