On Monday, April 23, 2001 at 9:30:21 PM UTC+5:30, Robin Winsor wrote:

I'm using the "Display equation on chart" option within the trendline function to fit a polynomial curve and see the resultant equation. The reason for wanting the equation is that I have two curves which I want to multiply. To do that I need the equation so I can replot them with evenly spaced datapoints. So here's the problem:

Excel gives an equation which, when used to construct a curve, does not
match the trendline drawn on the chart. Its close for much of the curve but it goes unstable at the end. The higher the order of polynomial the worse
it gets. Does anyone know why? I am wondering about rounding errors or something of the sort. If that is the problem, I wonder if there is a way
to increase the accuracy of the coefficients. If thats not it I can only think it must be some sort of bug although I've never actually seen a real bug in Excel's calculations before.

I'd appreciate any help or thoughts on the problem.

Robin

Thanks Dave. I was having similar problem while fitting a 5th order polynomial on covid data. I stumbled upon this discussion. I should have checked this aspect given that Microsoft uses every item as an object model which can be independently accessed.

Jibanananda Roy

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)