FORECAST.ETS produces different result to exponential trendline on scatter graph

Hi there,

Is anyone aware of the difference between the calculation behind the exponential trendline function in an Excel scatter graph and the FORECAST.ETS formula?

They are providing different results with the same data set and I'm not sure why..

Thanks!

ETS stands for "Exponential Triple Smoothing" which is a smoothing algorithm that is completely different from the least squares regression algorithm behind the exponential trendline (and the LOGEST() and GROWTH() functions). The chart trendline is a based linear regression to the equation y=b*exp(mx) [linearized as ln(y)=ln(b)+mx]. ETS is a signal smoothing algorithm that attempts to take out the noise in a signal to get a smoother signal (Wikipedia description: https://en.wikipedia.org/wiki/Expone...tial_smoothing ).

If you are trying to replicate the chart trendline in the spreadsheet, use the LOGEST() function (which regresses against the equivalent y=b*m^x) or use the GROWTH() function.
https://support.microsoft.com/en-us/...rs=en-us&ad=us
https://support.microsoft.com/en-us/...6-21324e68b80d

Also note, depending on the set up, FORECAST.ETS will either auto detect seasonality in forecast (based on data set) or use preset seasonality in forecast.

This will be drastically different from linear regression, as regression does not take into account seasonal pattern.