Odd PR vs. SERPs issue

[earnest]

Dear members,

While at one of the webmaster forum, i came across an assumption which I must say took me a little aback. It goes basically as follows: there are no dramatic changes in SERPs except for just after a PR update. The changes in PR and SERPs are thought to take two different, yet somewhat simultaneous updates. So, the conclusion was made that you have to wait the 3 months it takes for an update to happen until you can see your backlinks affect the site's search results performance.

It seems really odd to me, as I have come across no indications of the above trend.

It would be great if you shared your own vision of this quaint thing.

[crankydave]

The toolbar is not a ranking metric. It does not affect the SERP's.

Dave

[wige]

That used to be true, somewhat. It sounds like the person is talking about the Google Dance, an old method of updating that Google used before switching to Everflux, which they use now.

Dave is correct, toolbar PR does not affect SERPs, however, updates in the PR database often used to coincide with SERP shifts.

Google Dance is a reference to the sometimes major changes in the SERPs that would occur after Google updated the underlying rankings algorithm. Often, these updates would be followed with the release of an update to the toolbar pagerank database.

Today, Google uses a technique they publicly refer to as Everflux, where changes to the algorithm are done incrementally, avoiding the big SERP shifts. However, smaller shifts can still happen if large directories get devalued, or if a factor that favored a group of sites is changed.

[Jeremy80]

Yeah I agree with wige. SERPs often change on a daily basis. SERPs use actual PR as part of their algorithm and that PR is constantly changing. The toolbar PR updating has nothing to do with SERPs.

[intelinside]

Exactly. toolbar PR takes time to get updated but SERPs will take into account the active or live PR rather than using the toolbar's PR.

And this is not an assumption, I have observed this on many occasions for my websites so I am pretty sure about it.

[earnest]

Thanks for the inputs...

Yes, it is quite clear that the toolbar pr has nothing to do with serps, but the thing I mentioned just worked very confusing, with a kind of blurry reference to actual pr updated roughly like the toolbar pr, i.e. not continuously, so i figured i'd better clarify the whole thing.

Come to think of it, how's the actual pr, the one partly accounting for serp fluctuation, updated (if the word "update" is applicable to the process)?

[wige]

Originally, pagerank was recalculated at the end of a crawl cycle in a recursive process known as normalization. This could have an effect on SERPs, except that I believe normalization now occurs constantly.

The old mechanism required that Google crawl the web for a certain (unspecified) amount of time. At the end of this time, lets say Google found 1 million pages and indexed them. Google would then go through all the links it discovered to figure out the odds of a user randomly selecting a specific URL - this is pagerank, the odds (x:n where n is the number of pages in the index) of someone randomly landing on the page. The higher the odds of that random surfer landing on a page that links to your page, the higher the odds they will then land on your page, and so on. Because this is recursive, and to take into account the fact that the index constantly increases in size, the entire database would periodically be normalized to account for the new number of indexed pages, and all pages would recursively get new pageranks.

My understanding is that new database technologies now allow that normalization process to be constantly running in the background - as soon as new pages are crawled and new links are found, the changes are immediately included into the index.

[earnest]

Oh, thank you, Wige, you seem to be launching an eye-opening process with me

Now that I think of it, Pagerank as the odds of a surfer landing on a particular page is an entirely logical perspective. I somehow got the wrong idea of it because the pervasive "importance of a page" reference is misleadingly vague.

Again, I appreciate your help.