Trendline

[Chas]

The Excel trendline equation generator is seemingly
incorrect in many instances - espectially at the higher
end polynomial fits... is there a fix for this problem?

 

[Tushar Mehta]

While it is possible that you have run into a problem with the
trendline generator, in more cases the problem lies in how one uses /
interprets the result. Maybe, if you share more information about the
problem that you have discovered...

--
Regards,

Tushar Mehta
www.tushar-mehta.com
Excel, PowerPoint, and VBA add-ins, tutorials
Custom MS Office productivity solutions

 

[Jon Peltier]

Chas -

The first thing you should try is increasing the number of decimal
places in the formula. For highest accuracy, use a scientific number
format with 15 digits.

- Jon
-------
Jon Peltier, Microsoft Excel MVP
Peltier Technical Services
Tutorials and Custom Solutions
http://PeltierTech.com/
_______

 

[Jerry W. Lewis]

What is your evidence for this assertion? With what data? In my
experience, when there are differences between the chart trendline and
LINEST, then the problem is numerically challenging and the chart
trendline (formatted to show full precision) is more accurate than LINEST.

Jerry

 

[Martin Brown]

The Excel trendline equation generator is seemingly
incorrect in many instances - espectially at the higher
end polynomial fits... is there a fix for this problem?

You would do well to check previous threads on this topic before making
unsupported assertions. If you can demonstrate a dataset where the Chart
trendline generates an incorrect polynomial fit of any order please do
so.

In all test cases that I have ever seen with difficult data the Chart
trendline is significantly *BETTER* at fitting a least squares
polynomial than the worksheet functions, data analysis or the basic
polynomial fit in certain other well know statistical packages.

I would suggest user error. Probably not displaying all the significant
digits of the polynomial coefficients at a guess.

Regards,