The Fractional PageRank Hypothesis

[Garrett]

Those who watch their PageRank closely may be interested in this purely speculative concept from the fine folks at SEORoundtable - fractional PageRank.

The concept emerged from this question - how does Google decide how to rank pages with the same PR? Though there could well be other factors involved, one hypothesis (only a hypothesis), from relaxzoolander of SEOChat, is that there are PR fractions that help Google decide which page should show up on top.

So if your site is PR5 how can you tell if it's a low PR or a high PR?

Digitalpoint, SEOChat poster (and creator of the Channel Data Charting AdSense Tool (http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040406NewAdSenseToolChartsChannelData.html)), postulated, "let's say you have a PR7 site, and normally everything one-click deep is PR6, well if you have a 'high' PR7, everything one-click in will also be PR7."

As rustybrick (http://www.rustybrick.com/) put it, "if your pagerank is currently a 6 and your main four inner pages are a pagerank of 5 then it is safe to bet that your homepage's PR 6 is a low PR 6."

Everyone involved in this discussion was careful to point out that their thoughts are purely hypothetical and that there are many many factors involved. I'll leave you with relaxzoolander's caveat to speculators, "perhaps someone knows some other information that will make this theory immediately 'void'."

SEORoundtable also presented an interesting thread (http://forums.digitalpoint.com/showthread.php?t=250&page=1&pp=10) on the value of links from forums.

Thanks to SEORoundtable (http://www.seoroundtable.com) for the tip and SEOChat (http://forums.seochat.com/) for hypothesizing.

[Mel]

Surely everyone knows by now that the labels on the Google Tool Bar are not real page rank, but handydandy labels?

Real PR could be a number between zero and a few billion. There is no need to calculate fractional PageRank nor is there any need to apply corrections if two pages happen to have the same PR as the other factors in the Google ranking algo are taken into account.

[cbp]

The hypothesis seems reasonable - certainly been seeing it. To get a PR4 is reasonably easy, a five is a bit harder and a 6 takes even more links - I do notice that on several of my sites that when they were PR5, they did not go a 6 when PR updated when the inner pages (one click from home) were still a 4 --> this implies that the PR5 was really a low PR5 and did not have enough links to go to a 6. When the inner pages got up to a PR5, I would assume that the home page was a high PR5 ---> sure enough it did not take much to push it over the threshold to a 6 at the next update. Seen this 4 times on 4 different sites now.

CBP

[janeth]

I agree with the fact that when the inner pages one click deep are the same as the home page pr that it is a high pr rather then a low one.

But I also agree with Mel


Real PR could be a number between zero and a few billion. There is no need to calculate fractional PageRank nor is there any need to apply corrections if two pages happen to have the same PR as the other factors in the Google ranking algo are taken into account.

[spidersam]

Hmmm. Interesting hypothesis. Makes sense I guess.

[Dave Hawley]

This theory would only hold water IF there are no other pages (internal and external) pointing to the page in question.