Help for LINEST/LOGEST is extremly misleading. They fit functions that are
linear in the unknown coefficients (not just straight lines) which includes
all chart trendlines.
For polynomials see
http://www.stfx.ca/people/bliengme/ExcelTips/Polynomial.htm
For logarithmic trendlines see
http://groups-beta.google.com/group...d1581/f3d0f2172c25a661?hl=en#f3d0f2172c25a661
For power trendlines see
http://groups-beta.google.com/group/microsoft.public.excel.worksheet.functions/msg/35c521d64df2f86a
For exponential trendlines see
http://groups-beta.google.com/group/microsoft.public.excel.charting/msg/d955cd4118991d45
For Excel versions prior to 2003, the chart trendline may be significantly
more accurate (assuming you format it to give scientific notation with 14
decimal places) than LINEST and or SLOPE/INTERCEPT. If function results
agree with the chart trendline coefficients, then that is probably good
enough. Alternately David Braden has posted code to read coefficients
directly from the chart (to displayed precision)
http://groups.google.com/[email protected]
and I have posted alternate formulas to calculate SLOPE/INTERCEPT to 2003
precision
http://groups-beta.google.com/group/microsoft.public.excel.worksheet.functions/msg/d6a03470e7a1c650
Jerry