In the original "anatomy" paper the following quote is found in the section explaining how pages are ranked:
Quote:
|
...Google counts the number of hits of each type in the hit list. Then every count is converted into a count-weight. Count-weights increase linearly with counts at first but quickly taper off so that more than a certain count will not help...
|
Largely because of this and the fact that keyword densities above a certain value do not appear to affect Google rankings I have always assumed this system makes the calculation of keyword percentages redundant.