webdude - 5:55 pm on Aug 12, 2004 (gmt 0)

Huh? That is precisely the point of this type of experiment, to keep other results from contaminating your petri dish...

I was refering to his comment...

So the ultimate optimizing may be file name, H1, bold, underline, and italics, then meta tags. The more the better. Who cares about a PR and back-links!

How can you use these results in a real-world scenario? What he is doing may give the results of how to display and rank unique text, but not how to rank with others. PR and backlinks are two of the determining factors for this. If you take some of the factors in determining page rank and throw out the others, the test is a moot point.

However, if you are using the experiment to just see the value that the G places on text and file layout, then maybe he has a point.