That was quick Cap'n! I shall peruse your return zip.
Remember, all I did was take 4 x,y pairs, and make a best fit formula of straight line type.
Also, that was for only one sample, a straight line might not be best for the other 220+ samples, as you found out!
Mind you we don't know how far you can extrapolate/interpolate the data?
EDIT:
Sorry, can't unzip the file, its only 22 bytes?
N.