Trendline and linest gives incoherent results

Thank you davsth for your quick reply,

Well it is interesting to see that other codes give same results, but still when I plot the polynomial function I get a crazy plot(grey curve in the plot)...
Attachment 631604

EDIT: I've used coef. given on the web site you shared. It gives the good correlation. But it makes the analysis of my data on excel very tedious. Can I get the trendline coef. via VBA?

The problem I see is that, when you copied the trendline coefficients into the spreadsheet, you did not copy the coefficients to their full precision. for example, your x^6 coefficient is entered with only 4 significant figures -- -5.902e-10 The trendline shows this coefficient as -5.90204727382485e-10. I expect that, if you will copy the trendline coefficients into the spreadsheet to their full precision, you will get adequate results.

The LINEST() coefficients seem to give an adequate fit of the data as well. LINEST()'s colinearity algorithm appears to have kicked in and decided that the x^1 term was colinear with one of the other terms and it forced the x^1 coefficient to 0, which necessarily changes the other coefficients, but it still gives and equivalent fit. Why do you prefer the chart trendline over the LINEST() trendline? I would also be curious if there is a way to override the colinearity algorithm so that it will give the same coefficients as the other regression algorithms.