TREND in WPS Spreadsheet: Predict Values from a Linear Fit

For x values 1,2,3 and y values 5,8,11, =TREND(B2:B4,A2:A4,4,TRUE) has an independent expected result of 14. The fitted line is y=3x+2. TREND returns fitted or predicted y values; it does not return the slope and intercept as its main result.

TREND(known_y's,known_x's,new_x's,const)

known_y's is the required numeric response dataset. known_x's is optional; omission assumes 1,2,3,… matching the y dataset’s size. new_x's is optional and defaults to known_x’s. If both x arguments are omitted, both use that implicit sequence. const is an optional logical value: TRUE or omission fits an intercept normally, while FALSE fixes the intercept at 0 and refits the other coefficients.

With one predictor, known x and y must have matching dimensions. With several predictors, known y must be a row or column vector. Each independent variable gets a separate column when y is a column, or a separate row when y is a row. New predictors must preserve that variable arrangement and count.

Enter x=1,2,3 in A2:A4 and observations 5,8,11 in B2:B4. Enter 4 in D2. Use =TREND(B2:B4,A2:A4,D2,TRUE) in E2. Expect 14 under the documented model. With D2=5, expect 17.

An independent calculation gives slope (11−5)/(3−1)=3 and intercept 5−3×1=2. Every observation lies exactly on that line. For noisy data TREND instead finds a least-squares fit; it does not have to pass through each observation.

For these same data, forcing the intercept to zero changes the slope to Σxy/Σx² = 54/14 = 27/7. Therefore:

=TREND(B2:B4,A2:A4,4,FALSE)

has an independent expectation of 15.428571428571, not 14 and not 12. FALSE causes a new fit, rather than deleting the old intercept after fitting. Only use this constraint when a zero response at x=0 is justified by the application.

=TREND(B2:B4,A2:A4) returns fitted values at the original x values, independently expected as 5,8,11. A multi-value output requires the correct array-entry or spill behavior for your edition. The historical WPS lesson describes array formulas; it does not establish universal current spilling behavior.

WPS’s source describes using columns for x, x² and x³ to fit a polynomial through a linear regression in those predictors. Keep the same powers in new_x’s, enough nonredundant observations and a realistic interpolation range. Such a fit is still linear in its coefficients; it is not automatically safe for long-range forecasting. Use LINEST when you need coefficients rather than predicted y values.

A wrong but plausible result often comes from reversing known y and known x, misaligning rows, or leaving new_x’s blank when you intended a future period. Check all three ranges before changing the model. A single displayed value may be only the first element of an array result. Historical Ctrl+Shift+Enter instructions should be applied only where supported by the installed build.

The walkthrough’s TURE spelling is a typo; the logical constant is TRUE. It also describes the const option loosely: the formal contract fixes the intercept at zero when FALSE, not the slope. Unlike exponential GROWTH, a linear model can represent negative responses. Forecasts remain conditional mathematical outputs, not guaranteed future values.

The official WPS TREND lesson was checked on October 19, 2026. Its edition banner identifies WPS Office for Windows Personal. This Chinese first-party documentation is not proof of universal English-global, Mac, Linux, web or mobile support. Introduction and retirement versions remain unknown. All numerical results in this guide are independently calculated expectations, not native WPS test results or borrowed Excel outputs. Current GUI access was unavailable; no current-application screenshots or native execution are claimed.

FORECAST, INTERCEPT. Browse the WPS function library.