Which you can do with LINEST etc. or more directly with TREND or
FORECAST by including just the relevant surrounding points (instead of
the entire data set) for a given interpolation.

Jerry

MrShorty wrote:

I would agree with the OP on this one. LINEST etc. are curve fitting
functions, not interpolating functions. Sometimes when you have a
table of data, it is easier/preferable to approximate the function as a
series of straight lines connecting the data points (similar to what a
line chart/XY chart does) rather than attempt to fit the data to a
curve.