Truth or Consequences: Leaky PR?

[mjtaylor]

Does PR leak from your site or your home page when you have outbound links? I have heard some SEOs say that internal PR is calculated separately than outbound PR, as it were.

For example, if PR can leak, how do directories keep theirs?

What do you have to say about this?

MJ

[crankydave]

Basically, it's an indirect PR "loss or leak".

All pages have an available amount of PR to pass onto pages they link to.

Anytime there's an OBL off site, the amount of PR that can be passed internally by links on that page is lessened.

Dave

[SemAdvance]

It does not leak!!!

It passes!!!

All sites pass Page Rank with the outbound links the site has.

No leak involved....

If sites did not pass PageRank via OBL, then there would be no PageRank assigned to any site.

So called SEOs like to spread myths and fallacies, so others who try to understand SEO won't....

[ladraut]

I know that many will disagree, but pagerank is passed, not leaked.

There is no such thing as pagerank "leakage". A Web site or page doesn't LOSE pagerank with outbound links, it simply passes a theoretical percentage credit of the page's pagerank.

At least that is how the Google guys orignally envisioned it.

In practice though, it is not based on the number of links on the page, either site navigation or outbound. I can hear the objections screaming at me now.

But here is a clear example:

This past January, I had a brand new site, with only ONE link to it (from a PR5 of mine with 15 navigation and about 20 outbound links on this homepage that I linked from), and the home page of this new site became a PR3.

Google isn't revealing how they credit "link juice", and I'm sure that they laugh everytime an expert quotes the formula from the Google guys' paper from their university days.

[espmartin]

Passed, not leaked...

Is this an argument over the semantics of the two words,
in relation to PR and how it flows? Or are we discussing the
terms by their definitions?

Oh, I threw out yet another term (PR flow) - please don't flame me

[Webnauts]

Why shouldn't we all not have a look at this article and then continue the discussion:
Inbound and Outbound links

[deepsand]

Originally Posted by espmartin

Passed, not leaked...

Is this an argument over the semantics of the two words,
in relation to PR and how it flows? Or are we discussing the
terms by their definitions?

Oh, I threw out yet another term (PR flow) - please don't flame me

The word "leakage" implies a loss, in this case, a loss of PR for the referring page, which is, in fact, not the case.

While the word "pass" does not imply a loss, it does imply a transfer, which is also not wholly correct.

Absent a single word quickly springing to mind which accurately describes the process, let it be thought of as the case that the referring site creates PR, the amount of which is determined by its own PR, and then gives said created PR to the referred to site.

[DoneInStyle]

Why shouldn't we all not have a look at this article and then continue the discussion:
Inbound and Outbound links

That all sounds like someone really thought it out, but WHO? Who is Phil Craven? Does he work for a Search Engine? Who declared that he's a "PR Expert"? Him? His wife? His dog? Who developed this pagerank formula? How has it been verified?

In my experience, pagerank passes. It doesn't diminish on the page it's being passed from.

[Webnauts]

Has the term PageRank Dilute appear yet in this thread? Or is that only about internal pages?

[espmartin]

Great read. Along those lines, and hopefully not too far off topic, what I've done with
nofollow is to "take control of PR flow". I've used nofollow on internal pages to "flow" PR to
pages I deem important for the site's credibility. I've nofollowed my contacts / about /
accessibility statement / etc" - and in theory, this should allow the existing and incoming PR
to "flow" to my other pages.

[Webnauts]

Originally Posted by DoneInStyle

That all sounds like someone really thought it out, but WHO? Who is Phil Craven? Does he work for a Search Engine? Who declared that he's a "PR Expert"? Him? His wife? His dog? Who developed this pagerank formula? How has it been verified?

In my experience, pagerank passes. It doesn't diminish on the page it's being passed from.

I did not mean that I agree with him. I suggested all to have a look and then bring their comments here. And remember I have not expressed my opinion or facts yet.

[mjtaylor]

Originally Posted by ladraut

I know that many will disagree, but pagerank is passed, not leaked.

There is no such thing as pagerank "leakage". A Web site or page doesn't LOSE pagerank with outbound links, it simply passes a theoretical percentage credit of the page's pagerank.

According to the article to which Webnauts referred, PageRank, as it is passed does reduce the PR of the site.

Originally Posted by ladraut

At least that is how the Google guys orignally envisioned it.

In practice though, it is not based on the number of links on the page, either site navigation or outbound. I can hear the objections screaming at me now.

But here is a clear example:

This past January, I had a brand new site, with only ONE link to it (from a PR5 of mine with 15 navigation and about 20 outbound links on this homepage that I linked from), and the home page of this new site became a PR3.

Well, yes, but a site on its own does have PR - as I (thought) I understood it, any given page is presumed to have a PR1 ... so a site depending on how large it is will start out with PR despite few inbound links.

Originally Posted by ladraut

Google isn't revealing how they credit "link juice", and I'm sure that they laugh everytime an expert quotes the formula from the Google guys' paper from their university days.

Originally Posted by Webnauts

Why shouldn't we all not have a look at this article and then continue the discussion:
Inbound and Outbound links

So, Webnauts, that article clearly states that PR is leaked when links are outbound.

If that's true, then how do large directories maintain their PR?

I think it's quite likely that ladraut may be right - that PR is not calculated as it once was - and that the leaks may have been plugged up, so to speak.

Any more definitive answers to this out there?

Cheers, MJ

[deepsand]

Originally Posted by Webnauts

Why shouldn't we all not have a look at this article and then continue the discussion:
Inbound and Outbound links

Mr. Craven's notion that OBLs result in a loss of PR to the referring site leads to a mathematical conundrum, owing to the fact that such is infinitely recursive.

To explain, let us assume that a given site initially has a total PR amount of 1, that the amount of PR passed to via an OBL is half of its total PR, that it has but 1 OBL, that the inital PR of the referred to site is 0, that the referred to site has no OBLs, and that these are the only 2 sites indexed.

On the 1st iteration, the referring site loses a PR amount of 0.5 via the OBL, leaving it with a total PR of 0.5 remaining, while the referred to site now has a PR of 0.5.

On the 2nd iteration, the referring site, now having but 0.5 of PR, passes only 0.25 of PR, leaving both it and the referred to site with 0.25 PR each.

On each successive iteration the PR of each site is reduced by half, so that the total combined PR of the 2 sites approaches 0.

Where did the missing PR go?

[crankydave]

Originally Posted by espmartin

Passed, not leaked...

Is this an argument over the semantics of the two words,
in relation to PR and how it flows? Or are we discussing the
terms by their definitions?

Oh, I threw out yet another term (PR flow) - please don't flame me

LOL! And yes.

It's little more than a semantic discussiion.

The amount of PR that can be passed/flowed/whatever within a site internally is lessened by the external links.

Ergo... The amount of PR that can be passed internally is leaked/lessened/lost/whatever by the external links that pass PR.

Gawd... I love how folks scream myth/bogus/lie/otherwise. PR that is passed... INTERNALLY... is (repeating myself) INDIRECTLY "lost or leaked" by external links due to the amount of PR that CANNOT be passed by the remaining INTERNAL links.

Dave

[mjtaylor]

Originally Posted by crankydave

LOL! And yes.

It's little more than a semantic discussiion.

The amount of PR that can be passed/flowed/whatever within a site internally is lessened by the external links.

Ergo... The amount of PR that can be passed internally is leaked/lessened/lost/whatever by the external links that pass PR.

Gawd... I love how folks scream myth/bogus/lie/otherwise. PR that is passed... INTERNALLY... is (repeating myself) INDIRECTLY "lost or leaked" by external links due to the amount of PR that CANNOT be passed by the remaining INTERNAL links.

Dave

So, explain to me, please, how large directories keep their PR .... they are "leaking" PR all over the place ...

And I do not mean this to be a discussion of semantics.

I want to know if PR is truly divided among all links from a page, regardless of the link's destination ...

[Webnauts]

I suppose that it is clear here, that quality outbound links have positive effects within other parts of Google's ranking criteria. And PageRank is not the only ranking criteria. Having an outbound link, PR might pass/leak/lose/flow or jump in a river PR, but through such implementation you will meet another ranking criteria or requirement, which may have more value or equal? I do not want to go further in that though.

And MJ can you tell an example of a directory that does not have IBLs, and only OBLs and has PR or ranks well?

[deepsand]

As regards Mr. Craven, will no one address the unescapable mathematical issue that I've noted?

If you cannot logically, mathematically explain where the PR goes in the example given, then you cannot logically hold that PR is "lost" via OBL's, unless you can posit a PR "flow" that differs from that portrayed by Mr. Craven's article.

This is not to say that he is necessarily wholly wrong; consider that his portrayal may be correct, but incomplete.

Points to ponder:

1) How is PR created?

2) Within a fixed universe of static sites, is total PR conserved?

__________________________________________________ _______

[Webnauts]

Originally Posted by deepsand

1) How is PR created?

I would like to pose the question otherwise: Is PR based only on links?

[deepsand]

Originally Posted by Webnauts

I would like to pose the question otherwise: Is PR based only on links?

Were that the case, then it should be clear that there is no escape from the aforesaid "evaporation" of PR via infinite descent.

Still, allowing that IBLs are not the sole source of PR does not escape the the question posed re. conservation of PR in a static universe.

__________________________________________________ _______________________________________a

[Webnauts]

Originally Posted by deepsand

Were that the case, then it should be clear that there is no escape from the aforesaid "evaporation" of PR via infinite descent.

Still, allowing that IBLs are not the sole source of PR does not escape the the question posed re. conservation of PR in a static universe.

__________________________________________________ _______________________________________a

Does this answer the question? Google Technology Or did I get the question wrong?

[deepsand]

Originally Posted by Webnauts

Does this answer the question? Google Technology Or did I get the question wrong?

No, it does not answer the question. Note that the cited source says that links are counted as "votes," but says nothing to the effect that the "voter" loses anything by casting a vote.

And, note that the point to ponder re. conservation of PR is set in a fixed universe of static sites; i.e, the number & identity of the "voters" remain unchanged, as do their "votes."

Now, if PR "flows," i.e. moves from voter to another each time the votes are counted, in the manner described by Mr. Craven, such that the weight of the votes changes with each count, with the mathematically inescapable result that the total weight of the votes decreases with each counting, eventually all votes become worthless; i.e., PR is not conserved.

__________________________________________________ _______________________________________

[bj]

That Google page says that PR is similar to votes that are WEIGHTED depending on where the incoming links come from. It says nothing about any loss by the site that casts the vote.

And my niche directory, which is nothing BUT outbound links, ranks just fine.

I think anyone who thinks that Craven is right is not analyzing the reality that exists.

[DoneInStyle]

According to the article to which Webnauts referred, PageRank, as it is passed does reduce the PR of the site.

I just read that article three times slowly and didn't see ANYTHING that says PR is reduced, or even hints at it. Please tell me what you read that you interpreted that way.

[cass-hacks]

Originally Posted by Webnauts

Why shouldn't we all not have a look at this article and then continue the discussion:
Inbound and Outbound links

I've a few problems with the linked to article.
First, this statement :

Outbound links are a drain on a site's total PageRank. They leak PageRank.

is made as if a fact yet with no supporting evidence of it being so. One is expected to take the word of the author for it being so without even an attempt an an explanation.

I wonder why it is that the author tries to go into such detail in other areas yet seems to totally gloss over any explanation as to exactly how the link-juice drainage is supposed to occur.

Second, why is the author worried about an action attribute of a form element? Why would a search engine even care?

Third, Google's original patents concerning PageRank calculations as well as white papers written on the subject don't seem to mention anything about external linking ending up causing a loss of link-juice. The formula most often quoted doesn't have a diminishing term anywhere in it so how can a page end up with less than it started?

On the other hand, if one considers all links together, both internal and external and considers them all to share in the PageRank a given page has to pass around, then it stands to reason that PageRank divided equally among internal and external links on a given page will see the internal links receiving a smaller share than they would had there been no external links.

Although I may just be quibbling over semantics, I wouldn't call that a PageRank drain but yet I would agree that external links on a page reduce the PageRank available for internal links on the same page.

But, if the number of external links on a given page is small in proportion to the internal links, any "loss" of PageRank likely would be "minimal".

[inertia]

Im confused! If PR didnt transfer through links then PR would be a useless concept? An inbound link is an outbound link somewhere else!

So therefore outbound links do carry PR? Is this a debate about that or is this a debate about whether the PR is reduced on the page with the OBL? If thats the case then explain this - i managed to get a link on a PR 7 page to my home page, within 1 month my home page PR jumped from 2 to 4 and the SERP for the anchor text shot from page 5 to page 1. Now the page rank has obviously jumped from 2 to 4 because of this link, but the page rank on the PR7 page is till PR7! Surely if it has passed the PR on to my page it would have dropped?

[cass-hacks]

Originally Posted by inertia

or is this a debate about whether the PR is reduced on the page with the OBL?

Yes, no, maybe.

The debate is about PageRank "leakage" due to outbound links. The question that needs to be answered before there can be any debate though is what is the "leakage" that supposedly takes place and then how does it happen.

Originally Posted by inertia

If thats the case then explain this - i managed to get a link on a PR 7 page to my home page, within 1 month my home page PR jumped from 2 to 4 and the SERP for the anchor text shot from page 5 to page 1. Now the page rank has obviously jumped from 2 to 4 because of this link, but the page rank on the PR7 page is till PR7! Surely if it has passed the PR on to my page it would have dropped?

First of all, basing any conclusion or discussion on Toolbar PageRank is almost worthless.

Second, if Google's internal PageRank is logarithmic as many people think, which I wouldn't argue with, the difference between a change from 2 to 4 is much smaller than a change from 6 to 7.

For example a change of 2 to 4 logarithmic would be a change of 10 squared to 10 cubed which would be 9,900. But a change of 6 to 7 logarithmic would be 9,000,000.

That's not to say I believe the PageRank of a given page with outbound links would be diminished due to the outbound links themselves but instead, your example doesn't really prove the point you are trying to make.

On the other hand, if the PageRank a given page has to pass is divided among all the links on the given page, both internal and external, the more external links there are as compared to internal links, the less of that page's PageRank will go to the internal links.

You divide a pie into a number of pieces but you still have only one pie. The pieces of pie you give to others, the fewer pieces you keep for yourself.

[crankydave]

wow...

Every page has a finite amount of PR that can be passed.

PR is passed via the links on a page.

Any page that has a link on it that passes PR OFF SITE, reduces the amount of PR that can be passed INTERNALLY by the remaining links.

The amount of PR passed by that link that points off site CANNOT be passed by the internal links.

A site that has ZERO links pointing off site passes the total available PR it has back to itself.

A site that has ONE link that passes PR OFF SITE reduces the total amount of PR that is available by what ever it passes. That amount of PR is what is referred to being "lost/leaked/not passed/whatever" you want to call it.

Deepsand... Mathmatically it is a simple equation. If a page has a total available PR value of "x" to pass, and a single outbound link that passes PR reduces that amount by "y", then the available PR the page has to pass internally by the remaining links is "x-y".

MJ... As far as directories are concerned, all of them have internal links so they do keep some of the PR that is available to pass internally. Do not forget that internally passed PR is only part of the equation. External pages that link to the directory pass PR.

Look folks, I don't care what you want to call it, ANY link that directs ANY portion of PR away from a site reduces the amount of PR that can be passed INTERNALLY.

Dave

[deepsand]

Originally Posted by crankydave

<snip>

Deepsand... Mathmatically it is a simple equation. If a page has a total available PR value of "x" to pass, and a single outbound link that passes PR reduces that amount by "y", then the available PR the page has to pass internally by the remaining links is "x-y".

<snip>Dave

Okay ... so far. Now, repeat the process, a 2nd time, a 3rd time, ad infinitum.

Remember, the process is re-iterative; i.e., the "voting" is repeated, with the weights of the "votes" in each "election" being dependent of the results of the previous "election."

__________________________________________________ __________________________________

[crankydave]

Originally Posted by deepsand

Okay ... so far. Now, repeat the process, a 2nd time, a 3rd time, ad infinitum.

Remember, the process is re-iterative; i.e., the "voting" is repeated, with the weights of the "votes" in each "election" being dependent of the results of the previous "election."

__________________________________________________ __________________________________

Exactly. And with each iteration, the amount of PR that is being passed by that external link cannot be passed internally.

I really think folks get hung up on this way too much. While the "direct" result of an OBL does cause PR to (lost/not be passed/leaked/whatever) internally, the end result (actual PR assigned to a page) after all interations on all pages, can be positive, negative or neutral depending upon where the link ultimately leads.

For anyone still scratching their head, all a "PR Leak" refers to the direct loss of any amount of PR that can be passed internally due to OBL that passes PR. That's it.

Dave

[deepsand]

Originally Posted by crankydave

Exactly. And with each iteration, the amount of PR that is being passed by that external link cannot be passed internally.

I really think folks get hung up on this way too much. While the "direct" result of an OBL does cause PR to (lost/not be passed/leaked/whatever) internally, the end result (actual PR assigned to a page) after all interations on all pages, can be positive, negative or neutral depending upon where the link ultimately leads.

For anyone still scratching their head, all a "PR Leak" refers to the direct loss of any amount of PR that can be passed internally due to OBL that passes PR. That's it.

Dave

To repeat, "the process is re-iterative; i.e., the "voting" is repeated, with the weights of the "votes" in each "election" being dependent on the results of the previous "election.""

If you re-visit my example, I've already demonstrated the ensuing mathematical results, and how they lead to a total PR of zero remaining for all sites if Mr. Craven's portrayal were complete & accurate.

__________________________________________________ __________________________________

[crankydave]

Originally Posted by deepsand

To repeat, "the process is re-iterative; i.e., the "voting" is repeated, with the weights of the "votes" in each "election" being dependent on the results of the previous "election.""

If you re-visit my example, I've already demonstrated the ensuing mathematical results, and how they lead to a total PR of zero remaining for all sites if Mr. Craven's portrayal were complete & accurate.

__________________________________________________ __________________________________

Look... I said nothing at all about "Mr. Craven" now did I?

A page has a finite amount of PR to pass. Agree or disagree?

That PR is passed by the links on the page that pass PR. Agree or disagree?

The amount of PR that is passed is somehow divided among all the links on the page that pass PR. Agree or disagree?

ANY amount of PR that is passed to an external page CANNOT be passed internally. Agree or disagree?

Now, since you don't seem to like the phrase "PR Leak", you make up whatever phrase suits you for the amount of PR that does not DIRECTLY pass internally because it's being sent to an external page by an OBL.

As far as "Mr Craven" goes, you got a problem with him, take it up with him.

Dave

[crankydave]

Finally, where your "math" is concerned, since each and every page is given a fixed value at the start of each iteration and what is passed is added to that fixed value, no matter how many iterations you do the result will never be zero.

Dave

[martty]

Originally Posted by deepsand

The word "leakage" implies a loss, in this case, a loss of PR for the referring page, which is, in fact, not the case.

While the word "pass" does not imply a loss, it does imply a transfer, which is also not wholly correct.

Absent a single word quickly springing to mind which accurately describes the process, let it be thought of as the case that the referring site creates PR, the amount of which is determined by its own PR, and then gives said created PR to the referred to site.

Thats the best answer yet. The leakage thing is a myth

[crankydave]

Originally Posted by deepsand

The word "leakage" implies a loss, in this case, a loss of PR for the referring page, which is, in fact, not the case.

Yes it does imply a "loss". A loss of PR that can be transferred internally.

No, it does imply a loss of PR for the referring page. That is an incorrect "assumption" that gets made when the person making it does not understand what the "leakage" is referring to.

Dave

[deepsand]

Originally Posted by crankydave

A page has a finite amount of PR to pass. Agree or disagree?

Apparently not, if one allows that PR "flows" from one site to another.

Originally Posted by crankydave

ANY amount of PR that is passed to an external page CANNOT be passed internally. Agree or disagree?

Only if external and internal PR are part and parcel of the same thing.

Originally Posted by crankydave

Now, since you don't seem to like the phrase "PR Leak", you make up whatever phrase suits you for the amount of PR that does not DIRECTLY pass internally because it's being sent to an external page by an OBL.

As far as "Mr Craven" goes, you got a problem with him, take it up with him.

Apparently you subsribe to Mr. Craven's portrayal of PR as being a single entity; i.e., there is no distinction between internal & external PR. And, that is in fact the nut of the matter.

Simply claiming that such is the case, while ignoring the unavoidable mathematical consequences of such, will not suffice to make it so. Clearly the reality is not so simplistic.

By way of an analogy, do individual voters' powers decrease because they cast a ballot for President? Does the then elected President's authority diminish by virtue of his delegating authority to others?

__________________________________________________ __________________________________________________ _____________________

[deepsand]

Originally Posted by crankydave

Finally, where your "math" is concerned, since each and every page is given a fixed value at the start of each iteration and what is passed is added to that fixed value, no matter how many iterations you do the result will never be zero.

Dave

Really? Return then, if you will, to the analogy of an election.

If all voters vote for you, losing voting power with each election, eventually they will have no voting power remaining with which to re-elect you. What then?

Does the office that you held now stand vacant? Do you re-elect yourself in perpetuity?

While such might explain what oft times seems to happen in politics, it does not suffice to model peer-to-peer "voting" as employed in calculating PR.

__________________________________________________ __________________________________________________ _____________________

[crankydave]

Originally Posted by deepsand

Apparently not, if one allows that PR "flows" from one site to another.

Then you you don't understand PR or the definition of "finite". Regardless of the amount, it is finite.

Originally Posted by deepsand;343281}Only [u

if[/u] external and internal PR are part and parcel of the same thing.

There is no such thing as "internal PR" or "external PR'. PR get's passed internally or externally. If it get's passed onto an external page, it cannot be passed to an internal page.

If I have 5 "apples" and I give "Johnny" 2 of those "apples", how many "apples" can I give deepsand?

Dave

[deepsand]

Originally Posted by crankydave

Then you you don't understand PR or the definition of "finite". Regardless of the amount, it is finite.



There is no such thing as "internal PR" or "external PR'. PR get's passed internally or externally. If it get's passed onto an external page, it cannot be passed to an internal page.

If I have 5 "apples" and I give "Johnny" 2 of those "apples", how many "apples" can I give deepsand?

Dave

Invalid analogy. Voters do not give away their votes to the person they vote for.

__________________________________________________ ________________________

[crankydave]

Originally Posted by deepsand

Invalid analogy. Voters do not give away their votes to the person they vote for.

__________________________________________________ ________________________

I've really been trying to avoid your "voter" anaolgy with respect to "elections", since for lack of of a better way to put it, is really a pathetic anology and makes you look foolish to make the comparison between that and how PR is caculated.

Dave

[bj]

If I have 5 "apples" and I give "Johnny" 2 of those "apples", how many "apples" can I give deepsand?

This is ridiculous. Who determines the number of links on a page? We do. It's not something we have, it's something we bestow, either internally or externally. The only thing that limits the number of links we give is the amount of server space.

Quite frankly, I find this whole discussion rather foolish, since none of us work for Google nor know the algorithm. It's all guesswork. And I'm seeing some pretty funky guesses here, with absolutely no logic behind them. And Craven's wasn't any much better.

I'm outta here.

[deepsand]

Originally Posted by crankydave

I've really been trying to avoid your "voter" anaolgy with respect to "elections", since for lack of of a better way to put it, is really a pathetic anology and makes you look foolish to make the comparison between that and how PR is caculated.

Dave

Presumably that means that you consider Google's own explanation, at Google Technology , where they use the voter analogy, pathetic as well.

Interesting.

__________________________________________________ ___________________________________

[cass-hacks]

Originally Posted by deepsand

Presumably that means that you consider Google's own explanation, at Google Technology , where they use the voter analogy, pathetic as well.

Interesting.

As with all analogies, you can only take a given analogy so far.

Why not take the voter analogy even further and say that a site can't pass PageRank until it is of adult age?

Or, you only get one vote in each race so that means you can only have one link on a given page to a given type of site?

The reason why not is because those would be outside the scope of what the analogy was used for in the first place.

Originally Posted by bj

Quite frankly, I find this whole discussion rather foolish, since none of us work for Google nor know the algorithm. It's all guesswork. And I'm seeing some pretty funky guesses here, with absolutely no logic behind them.

We don't need to work for Google to know how PageRank works because the patent the covers PageRank is publicly accessible and according to Google Corporate Information: Technology

Google uses numerous factors including its patented PageRank™ algorithm

in other words, they didn't just write about it, they use it.

This isn't guesswork, it is simply a matter of reading what Google wrote as to how their system works.

[deepsand]

Originally Posted by cass-hacks

As with all analogies, you can only take a given analogy so far.

Why not take the voter analogy even further and say that a site can't pass PageRank until it is of adult age?

In fact, such is effected by way of including both a site's age & the duration of the Domain Registration as factors in the ranking algorithm.

Originally Posted by cass-hacks

Or, you only get one vote in each race so that means you can only have one link on a given page to a given type of site?

Again, implemented by way of attempts to defeat various means of link spamming.

Originally Posted by cass-hacks

We don't need to work for Google to know how PageRank works because the patent the covers PageRank is publicly accessible and according to Google Corporate Information: Technology
in other words, they didn't just write about it, they use it.

This isn't guesswork, it is simply a matter of reading what Google wrote as to how their system works.

Can you cite a statement by Google that unequivocally states that a site/page loses PR by linking to an external site?

[cass-hacks]

Originally Posted by deepsand

In fact, such is effected by way of including both a site's age & the duration of the Domain Registration as factors in the ranking algorithm.

True, neither has much to do with PageRank or the flow/passing/transfer/meandering/wandering of PageRank within or between sites. In understand you are not saying there is a connection, I just would hope people don't confuse the issues.

Originally Posted by deepsand

Again, implemented by way of attempts to defeat various means of link spamming.

????

Google limits valuing links to one link per given type of site???

Originally Posted by deepsand

Can you cite a statement by Google that unequivocally states that a site/page loses PR by linking to an external site?

Read Google's patent. There are numerous references to it around so you should be able to find the actual patent itself.

As far as unequivocally stating something, the "statement" comes from understanding the algorithm. There is no need or benefit to stating all the possible implications.

Besides, considering how people are known for exaggerating and taking things out of context that Google does say, e.g. links = votes, do you really think they would come out and say that links-out reduce PageRank available for links-in even though it is true and follows naturally from the PageRank algorithm?

All you have to do is use the published algorithm and do the math yourself. It isn't really all that hard.

[deepsand]

Originally Posted by cass-hacks

Read Google's patent. There are numerous references to it around so you should be able to find the actual patent itself.

All you have to do is use the published algorithm and do the math yourself. It isn't really all that hard.

Well then, if it's that simple & evident, why not show us?

Enquiring minds really do want to know.

__________________________________________________ ____________________________________

[cass-hacks]

Originally Posted by deepsand

Well then, if it's that simple & evident, why not show us?

Enquiring minds really do want to know.

You can find the patent at United States Patent: 7058628
which also includes examples of the formula applied.

[dartman]

Originally Posted by mjtaylor

Does PR leak from your site or your home page when you have outbound links? I have heard some SEOs say that internal PR is calculated separately than outbound PR, as it were. For example, if PR can leak, how do directories keep theirs? What do you have to say about this?
MJ

I'll take a #1 search result for a PR4 page over a #7 search result for a PR5 page anytime.
IMO there's too much emphasis on PR when better content is what helps jack a site to the top.

[crankydave]

Originally Posted by cass-hacks

Read Google's patent. There are numerous references to it around so you should be able to find the actual patent itself.

All you have to do is use the published algorithm and do the math yourself. It isn't really all that hard.

Originally Posted by deepsand

Well then, if it's that simple & evident, why not show us?

Enquiring minds really do want to know.

A page has an available PR of "X" to pass to other pages via links for the purpose of PR calculations.

That page can only do 1 of 4 things.

1. Pass all of its available PR to internal pages.
2. Pass all of its available PR to external pages.
3. Pass some of its available PR to internal pages and some to external pages.
4. Pass none of its available PR because it has no links.

Note... Passing PR has no direct effect on the assigned PR of the page itself. PR is calculated on the value of the links pointing to a page not away from it.

In the case of #4 above. All the links that are pointing to that page are removed from the PR calculation. In essence, it does not exist as far as the PR calculations are concerned. You can verify this by reading the patents.

Option #1 above... All of the available PR is being "voted/passed" back to pages within the site. The site does not experience any PR "leak or loss" of it's ability to pass PR back to itself for the purpose of the PR calculations.

Option #2 above... All of the available PR is being "voted/passed" to pages outside of the site. The site experiences PR "leak or loss" of it's ability to pass PR back to itself for the purpose of the PR calculations.

Option #3 above... Some of the available PR is being "voted/passed" to pages outside of the site. The site experiences PR "leak or loss" of it's ability to pass PR back to itself for the purpose of the PR calculations.

Remember, "voting" or "passing" PR for the purpose of the actual PR caculations has no direct effect on the PR assigned to the originating page.

The "loss or leak" of the ability to pass PR to internal pages because of the links to external pages can have an indirect effect of the PR assigned to the originating page that can be positive, negative, or neutral. Which effect it has is solely dependent upon where those links to the external pages ultimately lead.

Dave

[mjtaylor]

Originally Posted by crankydave

Passing PR has no direct effect on the assigned PR of the page itself. PR is calculated on the value of the links pointing to a page not away from it.



Dave

I don't have any sort of bend toward mathematics ... but if a Page A is passing all of its PR through internal links, and then an external link is added, a little less PR is passed internally, lowering the PR of the site's pages slightly. The PR of those internal pages passed back to the originating page would then be slightly lower, thus reducing (albeit indirectly) the PR of Page A.

That said, I have agree with BJ that "...none of us work for Google nor know the algorithm." So we don't really know whether internal and external PR is calculated independently or not.

I understand that there is a patent on a formula and that G uses it as part of their ranking - and that doesn't mean that the formula is applied today exactly as it was when the search engine was first created. In fact, I would say we can be pretty sure it isn't.

I appreciate the dialogue here (though it has gotten a little warm here and there) because it helps me think things through.

I have believed for a long time that putting a few good outbound authority links on a home page may lead to a boost in rankings and is worth the slight drain on PR, if, indeed, it *does* leak PR.

But that would be better discussed in another thread. So, Ive started one on whether outbound links on the home page are a good thing or bad thing.

Thanks for all the input,
MJ

[crankydave]

Originally Posted by mjtaylor

I don't have any sort of bend toward mathematics ... but if a Page A is passing all of its PR through internal links, and then an external link is added, a little less PR is passed internally, lowering the PR of the site's pages slightly. The PR of those internal pages passed back to the originating page would then be slightly lower, thus reducing (albeit indirectly) the PR of Page A.

Not neccessarily.

If that external link points to a page that points back to you, and both links are counted, then the indirect net effect *could* be a higher PR value for the site.

Dave