Forum Moderators: open
For example lets assume you own Page A. You have a backward link from Page B.
Now lets assume Page B has a Page Rank of 7.
Lets call the function Po(B) the number of outbound links of Page B.
My guess is...
If Po(B) < 20 it passes on PR6.
If 20 < Po(B) < 40 passes on PR5.
If 40 < Po(B) < 60 passes on PR4.
If 60 < Po(B) < 80 passes on PR3.
etc...
More generally... for Page B with a PR of X and N outbound links...
Passed on PR = X - (N/20) - 1
(division ignores the remainder since N is an Integer)
This formula is a bit of guess.
As you can see by my number of posts i'm a beginner so just wondered if you more experienced SEO's can improve it.
My assumption is that there is a linear relationship between outbound links and Page Rank however this is probably a wrong assumption.
Don't over simplify!
It is commonly accepted that PR isn't an integer and that it follows a geometric or even logarithmic progression.
Nobody but Google knows about the log base which is estimated around 6.
This means that a "strong PR7" (i.e. 7.999) could be up to 6 times stronger than a "weak PR7" (i.e. 7.000), and that the transmitted PR varies in the same proportion.
So, with a limiter number of outbound links, a strong PR7 could well transmit enough "votes" to give a PR7.
Dan
Yes, the toolbar PR is just for show, it is not the real one. Most of those who have studied it suspect that the toolbar PR is on a log scale, guesses about the base start at six and go quite a bit higher.
Assuming a base of six, a site which just reaches, say, PR5 on the toolbar will have one sixth of the real PR of a site which just fails to make PR6.
> Just how can you tell the difference between a strong and a week?
You cannot, Google has, intentionally or otherwise, truncated the amount of information it gives you.
Actually, for a home page that is listed in the Google Directory you can get a slightly better estimate, see the searchnerd site for details.
It follows from the mathematical expression e which can be normalized usining logarithms.
Basically the idea behind page rank is that google will always have a page rank of 10. an simple page with no links will have a page rank of zero .
in order to increase your page rank the amount of links needed also increases , between 8 , 9 and 10 this increases exponentially . This therefore allows for many low level page ranks and as the rank increases the amount of pages with this page rank decreases.
for example there 2 million pages with page rank 1 (they have 1 link ) and 2 pages with page rank 10 (they have 2 million links ) . So you can see how this changes everything .
but now page rank is not everything but links , because our phd friends are not so stupid . They have included the fact that keywords in links play an important role , that content is important and the fact that algortithm is googles entire business .
Now links can be broken down to crosslinking, interlinking, internal linking, external linking, canonicals, subdomains, directories, hubs, reciprocal, static links, dynamic links , so now we need to identify which links have more weight .
i would imagine external links have a weight of 0.7 while internal links have a weight of 0.5 . now 0.5 x 1000 links would equal a site that is huge however it would not get more than a page rank of 6 . why? beacuase internal links are weighted lower than external links . So we look at a one page site with 1000 external links that excellent i.e of page rank 8 or more and i bet that this site would have a page rank of 8 .
hope that helps
Presumably Google got its PR10 "honestly" because so many people link to them.
The formula is for calculating the amount of PR passed on from a backlink... in terms of the number of outbound links that page has and its PR.
A formula would be useful since it would be able to compare how effective a link from a particular site would be.
ie.)what page would pass on more PR... a backlink from a page which has PR4 and 2 outbound links or a backlink from a Page which has PR7 and 50 outbound links.
Also I never said that PR was an integer... I said that the number of outbound links is an integer which it clearly is...a natural number to be more precise.
Trying to make it simple. Let's think about "voting"
If we assume a base 5 log scale and that one vote is enough for a page to get a PR1 you'll have:
PR1 -> 1 vote
PR2 -> 5 votes
PR3 -> 25 votes
PR4 -> 125 votes
PR5 -> 625 votes
PR6 -> 3125 votes
etc...
You'll realize that, with anything between 625 and 3124 votes, a page will get a PR5 (PR6 needing 3125 votes)
As the "voting power" of a page is typically 85% of the amount of votes received, the same PR5 page will have a "voting power" between 531 and 2656.
The upper value (2556) is enough to "distribute" a PR5 to 4 pages.
Does that make sense?
Dan
