Hi all,

I am trying to fit a trendline to some data given by (x,y)-coordinates which I have plotted in a xy-scatter plot. I fit a exponential trendline which give a formula
  • 5.3234*exp(0.0061x)

and an R-squared value of 0.911. So far so good. This fitting looks very bad, actually. So I made a least square fitting using the solver and the grg nonlinear solver to optimize the parameters k and h in the expression
  • k*exp(h*x)

This lead to the expression
  • 36.0210863 * exp(0.003343605*x)
which yields a R-sqared value of 0.985. That is, a significantly better fitting, with the same function. Does any one of you know how excel fits these trendlines?
The data are
x y
100 3.569225062
200 17.77877873
300 63.61274436
400 89.51683479
500 169.8632096
600 292.3772838
700 410.8622885
800 568.3814671
900 756.7649326
1000 977.6198348