Trendline equation calculation

[Guest]

Equations generated by Excel for trendlines do not seem to match those
derived from the coeffiecients of a regression analysis.
And, when re-plotting data recalculated from the trendline equation for
original data, it is not the same - however when fitting a trendline to this
re-plotted data the same equation as the oringinal data is displayed!
Am I not writing the equation correctly, is this a genuine discrepency? Am
especially interested in polynomial fitted trendlines.
Many Thanks

 

[Tushar Mehta]

Right click on the equation in the chart, select 'Format Data
Labels...' then the Numbers tab. From the Category list select
'Number' and in the Decimal places field enter 15. This will give you a
much more accurate trendline equation.

 

[Guest]

Brilliant. Thank you very much. I found that increaing to about 6 decimal
places seemed to be accurate enough for me.

 

[Guest]

you can also use LINEST to get the coefficients into cells at max precision.
See http://people.stfx.ca/bliengme/ExcelTips/Polynomial.htm