As I recall, Braden's code also required that the graph be on a
worksheet, and would only extract the displayed precision.
Unfortunately, the chart trendline remains Excel's most stable and
accurate least-squares platform (hopefully that will change with the
next version of Excel). I would be glad to put your code through its
paces if you so desire.

Jerry

Tushar Mehta wrote:

Hi Jerry,

Do be aware that while Dave Braden's code might be the best option for
the job at hand, it has limitations. If I remember correctly, one of
them is that if a coefficient is 1, XL doesn't show it and the code
fails to handle that correctly. I have an improved version on some
computer but haven't shared it with the world because it has been tested
very lightly and, of course, XL2003 came along.