Calculating the number of links on a page

Will Google count what it's not allowed to see?

[yosmc]

As far as I know, if you put links on a page, page rank will be passed on according to the number of links on the page. The fewer the links, the higher the "value" of each link.

Now, how about those links that Googlebot isn't allowed to follow? If I have 20 links on a page, and 10 lead to a directory that is forbidden for bots (as specified in the robots.txt file), will that be 10 or 20 links on the page from Google's perspective?

[yosmc]

Anyone? :)

[chris]

i was wondering about that yesterday, i really want to know
the answer, because i'm currently trading links but if the other website in his robots.txt it's specified to skip
the links page what happens then.

[rogerd]

I haven't experimented with this, but my guess is that those would appear to be 20 links even though ten of them are in a forbidden directory. Chances are, GB just records the links it finds, and only checks robots.txt when it returns to get the page it found earlier. This isn't based on hard data, though, so perhaps someone else can comment with more authority.

[topr8]

i'm not sure of the answer,

but if you don't want google to count the links why don't you write them as
javascript ...

document.write instead

if you are trading links with someone who has banned googlebot from the links page you are wasting your time.

[yosmc]

topr8, in my case that would be a bit complicated. I have a message board with a lot of internal links on every page. Half of them are links to personal profiles, reply buttons etc. which are forbidden for the Googlebot to follow. Of course it would be better for my site structure & rankings if Google would ignore those links alltogether, but javascripting them would make it too complicated (apart from the fact that those folks who have Javascript deactivated would lose functionality).

[ciml]

Last time I checked (a few months ago) the PageRank handed on to the working links was diluted by the /robots.txt and 404 links.

On the other hand, two links to the same URL from the same page counted as one.

[figment88]

I also have no hard data.

The mechanics of the situation could very well be what rogerd hypothesizes, but it goes against Google philosohpy to allow non-followable links to dilute PR.

The pubicly available docs talk about normalization of PR which is kind of like the conservation of mass in the universe. You cannot send (or vote in their lingo) PR without a page target and expect the numbers to add up across all the pages in the index.

In addition, there are a lot of links (in the a href= sense) out there now that do not have another page at the end. There are all the "add item," "update information," "upload image" type of things that should have nothing to do with PR.

In a sense, the Google definition of a link would almost have to include the ability to indicate PR (i.e. a link is something that connects one page to a different page).

[ciml]

figment88, I had expected the same as you before finding the answer. For /robots.txt protected URLs it makes sense IMO; those URLs can appear in the results (although they are not fetched).

It does seem a little odd for 404s to count for dilution, but in terms of "conservation of mass" those URLs don't seem much different from pages with no links.