Full least-squares regression in one function: slope and intercept, and with stats TRUE the standard errors, R², F statistic, and residual sums. Handles several x columns for multiple regression.

Syntax

=LINEST(known_ys, [known_xs], [const], [stats])

• Slope and intercept

  =LINEST({2,4,6,9},{1,2,3,4}) → {2.3,-0.5}

  Slope then intercept, spilled across two cells.

• Pull R²

  =INDEX(LINEST({2,4,6,9},{1,2,3,4},TRUE,TRUE),3,1) → 0.988785047

  Row 3, column 1 of the stats block is R².

• Multiple regression

  =LINEST(D2:D50,A2:C50) → {b3,b2,b1,b0}

  Multiple regression on three x columns. Coefficients come back in reverse column order.

• The stats block is 5 rows: coefficients, standard errors, R² and standard error of y, F and degrees of freedom, regression and residual sums of squares.
• Coefficients are returned right to left: the last x column first, the intercept last.
• Wrap in INDEX to pull one number without spilling the whole block.

• #REF! when x and y have different lengths.
• #VALUE! when a cell is text or blank.
• Collinear x columns give near-zero coefficients with huge standard errors.