Page Rank Optimization - Additional Algorithms
(Page 3 of 4 )
Keep in mind that page rank was invented more that 10 years ago; Google has made modifications to the formula since then. Google also released Trust Rank, Topic Sensitive Page Rank (TSPR) and MANY other algorithms. We can assume that page rank is still one of the most important ranking factors, but don't build your entire campaign around this single measurement.
The Dampening Factor
Page rank is built on the principle of following links, resembling a random Internet user who follows all links he sees. The more links point to X page, the higher the chances of that random user finding the X page by following all the links he sees (hence links provide a clue to the importance of each page).
But what if the random user gets tired of following all the links and closes the window? This is the purpose of the dampening factor in the page rank algorithm. It comes in handy exactly at times when the user becomes "tired" and stops following the links (or stumbles onto pages that don't have any links to other pages).
According to Wikipedia, if a page has no links to other pages, it becomes a sink and therefore terminates the random surfing process. However, the solution is quite simple. If the random surfer arrives at a sink page, he picks another URL at random and continues surfing again. When calculating page rank, pages with no outbound links are assumed to link out to all other pages in the collection. Their page rank scores are therefore divided evenly among all other pages. In other words, to be fair to pages that are not sinks, these random transitions are added to all nodes in the Web, with a residual probability of usually d = 0.85, estimated from the frequency that an average surfer uses his or her browser's bookmark feature.
Pages Voting for Themselves
There is no actual proof as to whether pages can vote for themselves (i.e. URL site.com/page.html has link to site.com/page.html). The only thing we can do is guess that this sort of vote is impractical and is discounted by Google.
Page Rank Flow
If a page with PR 4 links out to eight pages, than each page that received a link gets 0.5 PR points. If that same page with links to only four pages, then each page gets 1 PR point. The distribution model is equal, meaning that points are equally divided between pages. The fewer links there are, the more page rank power is passed and conversely, the more links there are, the less PR is passed to all pages.
Not that pages do NOT lose page rank when they link out. Linking out merely distributes page rank values. For example, when a page with PR 5 links out to 10 websites, it does not lose its page rank, it simply distributes it, while keeping the actual value to itself.
More Than One Vote from a Page
We can assume that Google only counts one vote from a unique page and discounts duplicate or triple votes. For instance, if X page with a page rank of 4 has two links to A and one link to B, then the distributed page rank is 2 per each page, discounting the duplicate link.
This is an assumption. For further proof you can run your own tests.
Next: How Page Rank is Accumulated >>
More Search Optimization Articles
More By Ivan Strouchliak
/ 7 
