Koether:
Have you tried dumping the complete LINEST output, i.e. selecting a 4columnx5row region and using shift-ctrl-return to get all the coefficients and params?
And just for completeness' sake, try using the
'transpose' version posted above to see what you get.
I just ran a...
cub3bead:
You're absolutely correct - I've used CurveExpert for several years now.
An even better (more powerful) tool is Regress+, by Mike McLaughlin. However, that's MacOS only, which some of us regard as a bonus :-) .
I fully agree with the rule that Excel should be the platform of last...
Followup:
Several people on various microsoft/excel forums have told me this is a known bug in Excel2004 (OSX). The workaround is to use the transpose: (note the semicolon in there)
=LINEST(TRANSPOSE(y_range),TRANSPOSE(x_range)^{1;2},TRUE,TRUE)
The problem appears to be the way that...
There appears to be a bug in Excel2004 for MacOSX, 11.3.5 (at least). Here's an example of 2nd order polynomial fit. If I use the format
linest(yarray,xarray^{1,2},1,1)
the results are garbage. If I create two columns, one with x values and the next with x^2 values, then
linest(yarray...