Forum Moderators: open
Total of 3 PR sites linking to it with 2xPR3 & 1xPR4
When I've checked backlinks it just show two backlinks that PR4 external and the site itself.. PR3 links not showing as usual.
I wonder if that self backlink also affects PR? Like a 3rd party PR4 link?
Maybe if on www.domain.com/index.htm (or whatever the homepage is) you put a link on www.domain.com you might distribute some pagerank to yourself but I don't think this would be of great value.
Self backlink
Is that counting in PR?
I don't know the answer, but anecdotal observation makes me think the answer is yes, but it will seldom matter much.
I think this gets filtered out.
Otherwise there would be an endless loop.
Otherwise there would be an endless loop.
That's not correct. Both are valid (mathematical correct) models. There are no principle problems with self loops.
I haven't examined this problem so far, therefore I don't know the answer. (And you have so make a quantitative PR analysis to answer this question.) However, I noticed that the behaviour for the backlinks shown by the link command changed some months ago. Of course, this doesn't mean that the PR algorithm changed, but it might be a hint.
But would it make sense to count self-links?
But would it make sense to count self-links?
One have to look at the results - if the SERPs are improved than it makes sense.
Also, the question was if Google is counting these links for PR. They could do this even if it wouldn't make sense.
RE: www.webmasterworld.com links to www.webmasterworld.com
Aren't we talking about page "B" from site "A" linking to page "C" on site "A"? This may, or may not (cannot argue with that :o) count as PR. However, it would silly to have this "unknown" dictate which pages you link to on your site.
If the question really was www.webmasterworld.com links to www.webmasterworld.com (I still think it wasn't) then I would have to say no.
Yes, what I was asking if there is a kind of loop.
Let's say that domain is [domain.com...]
There is two backlinks:
[domain.com...]
[external-PR4.com...]
So when calcuating PR of that domain will google think as 2xPR3 & 2xPR4 back links?
If self-links can be shown as backlinks, it is quite logical that they do count for PR and anchor text...IMHO; no proof of it, only the smell in the air.
For example, anchor text with links pointing to itself seems to give some extra weight in relevancy & ranking higher than plain text.
Most PR<4 backlinks are not shown. With your logic they wouldn't be counted :)
I have just found a site with a PR6 but it only has backlinks from itself and only 32 of these. How can this be?
It has lots and lots of <PR4 inbounds.
TJ
So how many links of PR3 or less would it take to reach PR6, must be 1000's I would think.
If this kind of thing were as you are saying, I would imagine there would be a ton of pages linking to themselves multiple times and thereby multplying their PR.
We postulate that the Google algo can do all sorts of magical things yet we don't think these guys were smart enough to think of that?
So how many links of PR3 or less would it take to reach PR6, must be 1000's I would think.
Quite a few, coupled with a strategic internal link structure.
What you don't know, is whether PR and/or backlink data is up to date.
TJ
It doesn't mean that at all. Only one link counts, and yes it makes perfect sense for that to benefit the page, and for that benefit to be puny.
GodLikeLotus, I think you are confusing linking within a domain with a single page linking to itself, which is the issue here.
GodLikeLotus - see your sticky mail. Google works with pages not domains. Check the backlinks on sub-pages of that domain, not just the index page. The sub-pages have plenty of high(ish) PR inbound links, including a couple of PR6's. They all then point back to the index page.
TJ
Everybody is assuming all links are equal... for example text link at bottom of page vs. link at top of page vs. an image map vs. a link within H3 text etc. I am not sure we know they are equal to a fact for pr transfer do we?
Yes, I do.
I'm reffering to PR, not anchor text. Also, I'm not considering special cases as for example 1*1 images.
I would refer back to the original Brin paper. There seems to be a lot of misunderstanding about how PR is calc'd.
Apart from the fact that this question isn't answered in that paper, Google isn't using the original PR algorithm.
As already mentioned, both scenarios are valid mathematical models. Therefore, one has to analyze the current algorithm to find the answer.
For that purpose I have build some test pages, but I will take some time until I'll get a result.
For example, anchor text with links pointing to itself seems to give some extra weight in relevancy & ranking higher than plain text.
This neither proofs that PR is transferred nor anchor text is counted. The benefit might be just due to the change from plain text to the <a> tag.
Yet when I do a search like "blue widgets +a" My site is back up at #1. why would adding the word "a" make sucha huge difference, all sites use the word "a". Has there been any discussions on this, if so what do you guys think.
I didn't mean to discount that other on page factors (anchor text etc.) are likely more important factors than pr flow for linking to yourself say for an index page linking to an index page. Also, I agree that ten year old paper does not contain the current google algo... I would be shocked if it did. A good first go I would say that has ten years worth of hindsight applied to it by now.
Being totally unscientific here though... suppose I traded links with a like pr page. I get my incoming link at the top of their page, plain text, in bold or +1 text... and I gave a link at the bottom of my page in a 1x1 pixel. No one would do that because logic says if I prominently display a link in bold text at the top of my page people will likely see it and click on it... whereas the other link is just a link for google purposes. I just think that if that is common sense for us... google techs are at least as smart as we are.
I'm just saying that currently Google doesn't distinguish between different types of links for PR calculation.
Of course, there are cases where links are not counted as links, but there are no different types of links, i.e. if links are not ignored they have the same weight.
