Revisiting whether PR is lost when adding pages to a site [Archive]

bobmutch

04-11-2005, 11:47 PM

JohnW: "It is an opinion, and not one shared by many who know what they are talking about. The fact is that adding quality content to a site does not hurt PR unless it is done improperly."

No one said anything about hurting PR. You can view my series of diagrams of fully meshed and hierarchically structured models with a 40 iteration calculation and the source to the calculation on my site under the article heading "Page Rank Explained". (My site URL is beside my handle above.)

When you add a new page that has a real PR of 1 to a site there is going to be PR voted to that page on the next real PR update or if you set up a model and code you can see it real time. While the new page (it will have two way link or be an orphan) votes the real PR back some is retained. The retained real PR is "bleed" from the other pages hence the other pages decrease in real PR. This is easily noticed then you use a realistic model where there is inbound links to the home page.

I am not sure who you are referring to about those "who know what they are talking about." Could you produce some names and quotes.

I would suggest you consider Markus Sobek's section "The Reduction of PageRank by Additional Pages" in his article on "A Survey of Google's RageRank" article.

Here is a short quote for that section. "Since adding pages to a site often reduces PageRank for already existing pages, it becomes obvious that the PageRank algorithm tends to privilege smaller web sites."

Another well know name is Phil Craven that has written a good article "Google's PageRank Explained". Here is a quote from that article. "The new page will, of course, aquire PageRank from the site's existing pages. The effect is that, whilst the total PageRank in the site is increased, one or more of the existing pages will suffer a PageRank loss due to the new page making gains."

Both Markus and Phil are well known names and what I would consider PR experts.

JohnW it is been may experience that any one that does know what they are talking when it comes to PR for the most part, holds this position. It takes about 20 lines of code and 30 mins to prove it.

If you have a site structure where you think adding pages will increase the real PR of the home page and the main sub pages post it up and show where you propose to add the pages. I will quickly program up the code with 40 internations that will show the real PR of each page before and after the addition of pages.

I when through this once before with a couple of the regulars here on SEW that held a simular postion as you do. I posted numerous models, calculations and the code to show my point. They just keep on mumbling the same mantra and produced no examples and no code to back up their views.

I have yet to find a site structure that has inbound links to the home page where you can add pages in the site with out bleeding real PR from the home and or sub pages.

The idea that you can increase your site real PR may appear to work on a closed model where there are no inbound links. But as soon as you use a real to life model where there are inbound links it falls apart.

Ian Rogers idea of adding 1000 pages to a site with no inbound links only works on paper. As soon as you add enought links to get Googlebot to crawl those 1000 pages the small amount of real PR that is voted to the home page from those 1000 pages is nothing compared to the real PR that the inbound links vote to the home page.

The idea of going over 1000 links may work on paper also but with Googlebot only reading 101k you can't get much more than 1000 links on a page. Then you have to go to pages that are looped. So if you want to use 10,000 pages you are going to have to run 1000 links off the home page with a 10 deep loop.

Home to 1 to 2 to 3 to 4 to 5 to 6 to 7 to 8 to 9 to 10 to Home. With each page 1 -10 having a link to the home page.

The number of high PR links you will need to point at the home page to induce the Googlebot to read though those 10 deep loops will dwarf the small amount of PR you will get out of the 10,000 pages. And we have not even deal with the issue of orginal content.

I hold that the idea of producting PR by adding pages to a site is a myth and have yet to see any one produce a model (that will function in the real world) with code that works. In fact I havn't even met any one that holds the position you do take the time to try.

Marcia

04-12-2005, 02:08 AM

I am not sure who you are referring to about those "who know what they are talking about." Could you produce some names and quotes.Basically, who's being referred to is simply people who study and work out how to optimally structure the internal navigation of sites. There were a few who posted in those long threads on the topic a few months ago who do just that.

JohnW

04-12-2005, 08:28 AM

Interesting. First, you said:

>The fact of the matter is the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page.

Then, you said:

>No one said anything about hurting PR.

Anyway, let me repeat one thing I said before: if adding pages has hurt your PR then you are doing it wrong. This is starting to give me a déjà vu feeling, did we not put this subject to rest some time ago?

dannysullivan

04-12-2005, 08:54 AM

Since adding pages to a site often reduces PageRank for already existing pages, it becomes obvious that the PageRank algorithm tends to privilege smaller web sites.

-- and --

I have yet to find a site structure that has inbound links to the home page where you can add pages in the site with out bleeding real PR from the home and or sub pages.

So I disagree entirely. I'll offer up SEW itself as an example. I add lots of pages to it all the time and have done so for years. The PR of the home page has stayed steady between PR7-9 (currently at PR8). Adding content hasn't had any bad impact.

How about the SEW Forums? We have content, and lots of it, added every day. Home page has been steady at PR7 for several months, up from PR3 it had several months before that after launch.

Both sites have links back to the home page. In fact, I find it hard to think of many sites that don't link back to their home page.

How about Amazon? They're like a PR8 or 9 and have been that way as long as I can remember. They link back to their home page. They also add thousands of pages over the course of a month, conservatively speaking.

bobmutch

04-12-2005, 10:52 AM

JohnW: When you add pages to a site, the real PR will be voted to those new pages. The new pages, during a real PR update, will retain some of that real PR. That is just a fact. That is the way PR works. That is not to say that is a bad thing. It is just what happens.
I have never said that any thing "hurts" PR. I have just noted what happens. In fact to those that are not PR hoarders or trying to pump up there home page so they can sell links based on PR it is a good thing. You want there to be PR on all your pages. So I would say it is far from "hurts" PR but it is a natural occurrence that is a good thing.

"This is starting to give me a déjà vu feeling, did we not put this subject to rest some time ago?"
Yes John we did discuss this before. I show diagrams with calculations of the real PR before and after pages were added to differnet site models. It clearly showed what happens to the real PR on the home page and subs when you add additional pages.

These models seemed to be ignored and the mantra continues, "if your home page looses PR you are doing some thing wrong." The fact remains that most smaller sites are fully meshed via the menu system.

Calculation and common sense tells us that when you add a new page to a fully meshed structure all pages will decrease in real PR because a new page starts with a real PR of 1.

dannysullivan: "So I disagree entirely. I'll offer up SEW itself as an example. I'll offer up SEW itself as an example. I add lots of pages to it all the time and have done so for years. The PR of the home page has stayed steady between PR7-9 (currently at PR8). Adding content hasn't had any bad impact."
Yahoo shows over 424,000 IBL into SEW's home page so that is going to off set the bleed of real PR to your 60k or so pages. So all your example shows is that a huge number (424k) of IBL's to your home page will keep the real PR high and hence the toolbar PR high also.

"How about the SEW Forums? We have content, and lots of it, added every day. Home page has been steady at PR7 for several months..."
Again the home page of your forums sub domain is showing over 38k of IBL's in Yahoo which has been enough to keep that sub domains home page at a toolbar PR7 even though you have 19k or so forum pages.

Both these examples really only show us that if you throw enough links at your home page you can off set the real PR bleed to a large number of inside pages.

What you really needs to be done is to create a controlled model where and do the real PR calculation with a number of inside pages and with out those inside pages and see what the addition of the inside pages does to the real PR on the home page and the main sub pages.

This is what I have done and it’s pretty clear what the numbers show, at least with the original PR formula. Of course there is the issue that the original PR formula may have changed.

"Both sites have links back to the home page. In fact, I find it hard to think of many sites that don't link back to their home page."
Linking back to your home page is a good thing, both in navigation and in channeling real PR (during the real PR calculation of course) to your home page. If the objective is to have high PR (and I am not saying it is or should be) on the home page then of course put a link on every page to your home page.

"How about Amazon? They're like a PR8 or 9 and have been that way as long as I can remember. They link back to their home page."
I don't know where you are getting the idea what I or anyone has stated linking back to the home page is not a good think to do if the objective is to have high PR on the home page. Common sense tells us that is a good thing to do.

The issue is not about linking back to the home page. I issue that is being discussed as I see it is that the more pages you add to your site (with no change in the real PR vote to your home page by IBL's) the real PR will be voted and retained by those new pages during a real PR update. This will decrease the real PR on the home page in most cases, and will decrease the real PR on the home page and main sub pages in all cases.

I have not said that is a bad thing or "hurts" anything. It is just the nature of the original PR formula and with 30 minutes and 20 lines of code it can be clearly shown that this is the case - as I have shown in detail with diagrams, programs that run 40 iterations and a copy of the code for each example.
http://www.seocompany.ca/pagerank/page-rank-explained.html#number-of-pages

JohnW

04-12-2005, 12:02 PM

>The fact remains that smaller sites are fully meshed via the menu system.

If that’s the case, well, it goes to exactly what I said before – they are doing it wrong.

A fully meshed site navigation is not good SEO. (I hope I am not inviting another déjà vu moment, seems like I remember some crackpot ideas floating around on this as well…)

If we are talking about properly built websites, I still say you are missing the boat.

bobmutch

04-12-2005, 12:34 PM

JohnW: "If that’s the case, well, it goes to exactly what I said before – they are doing it wrong."
Personally I think fully meshed navigation is a good structure for smaller sites. It enables a person to go to any page from any page. Perhaps is it "wrong" if you objective is to channel PR to the home page. The only good reason I can think of doing that is if you want to sell links based on the PR of the home page or for bragging rights - "my site is a PR8". What benifits do you see with having higher PR on the home page as far as SEO goes?

"A fully meshed site navigation is not good SEO."
Why is fully meshed not good SEO for smaller sites? I would think that since real PR on longer holds much weight in the ranking algo, that whether the real PR of your home page is a bit higher or a bit lower will not make much differnet, if any, in your rankings and rankings is the main thing SEO is about.

I would think that if PR had weight in the ranking algo you would what it distrubited to the pages that you want to get better rankings on. Not to structure your site to try to channel as much as possible on the home page.
The idea that it's "good SEO" to channel your PR to the home page makes little sense to me. Like I said before, its important for those selling links on their home page.

"If we are talking about properly built websites, I still say you are missing the boat."
Missing the boat? What kind of a comment is that? I havn't even suggested or commented on how to built the structure of a website. I have only noted what happens with the PR vote on different site structures.

JohnW

04-12-2005, 09:26 PM

>What benifits do you see with having higher PR on the home page as far as SEO goes?

In addition to growing the little green bar, I think we can all agree that getting good links from good pages enhances a pages “ability” (lets call this the authority score or PR) to rank for relevant keywords, and, to affect the authority score of other pages it links to.

And what about the good page that links to it in the first place, what makes it a “good” page? The links that point to it, right? And so forth and so on. So it seems obvious that there is a authority score of some kind, PR, whatever you want to call it, and that it is good to have more of it and bad to have less of it.

Since you can control, though internal links and proper site design, how that authority is moved around within a site and concentrate it on important pages, that is what you would want to do as part of optimization. That’s also why earlier, I said full mesh navigation where every page links to every page, is bad SEO.

PhilC

04-12-2005, 09:59 PM

Since I was mentioned in this thread, I'll offer my view. To be honest, I haven't read all the long posts, so I don't really know who is saying what.

Adding pages to a site does increase the total PageRank within the site, but they will drain PageRank from the existing pages.

Take a 5 page site as an example. It has sufficient IBLs to give all of its pages some decent PageRank. It doesn't matter which pages receive the IBLs. There is a specific amount of PageRank within the site, and the PageRank of each page is determined by the linkages within the site. Now add 10 pages to the site without adding any additional IBLs, and link the new pages into the existing site. Each of the 10 new pages adds a little PageRank to the site as a whole, but they don't add anywhere near as much as each page already has. Suddenly, the existing PageRank (plus a very small amount for each new page) is spread around 15 pages instead of 5. The new pages have taken some PageRank from the original pages, and the original pages have less than they did before.

The fact of linking the new pages into the site, even if it's via just one link from any of the existing pages, means that the existing pages lose PageRank.

The exception to this is when the new pages don't link back to the originals. Any link to a page that doesn't link out is effectively dropped from the calculations - to all intents and purposes, the link and the page don't exist. So a new section with lots of new pages could be created and linked to, but if it doesn't link back, and if it doesn't link to anywhere else, no PageRank changes will occur within the original site.

So new pages that are incorporated into a site in a normal way, will drain PageRank from the existing pages as a whole. If it matters, then new IBLs are needed to compensate, or the linkages should be arranged to prevent the drain on certain pages.

The idea that you can generate toolbar PR from lots of internal links is a complete myth. It isn't a myth. With the right link structure it is easily done.

The fact of the matter is the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page.New pages take PageRank away from the existing pages, but not necessarily from the home page. With many or most link structures, the home page actually gains.

PhilC

04-12-2005, 11:02 PM

Here's an example of a 5 page site where the home page (A) links to all other pages (B, C, D, E), and all other pages link back to the home page. The site has 1 IBL. The extra pages (F, G, H) are not linked in and are completely ignored.

http://www.webworkshop.net/pagerank_calculator.php?lnks=2,3,4,5,13,25,37,49&ilnks=1&iblprs=&pgnms=Home+Page,,,,,,,,,,,&pgs=8&initpr=1&its=40&type=real

Note the PageRank values for each of the 5 pages:- 2.83, 0.75, 0.75, 0.75, 0.65

Now add one new page (F), and link it into the site in the same way as the other pages:-

http://www.webworkshop.net/pagerank_calculator.php?lnks=2,3,4,5,6,13,25,37,49 ,61&ilnks=1&iblprs=&pgnms=Home+Page,,,,,,,,,,,&pgs=8&initpr=1&its=40&type=real

The PageRanks have changed to 3.29, 0.71, 0.71, 0.71, 0.71

The home page gained and the other pages lost. More pages can be added, and linked into the site in different ways, but the existing pages as a whole will always lose when new pages are added and linked into the site in a normal way, even though one or more of them can be engineered to gain.

bobmutch

04-12-2005, 11:32 PM

PhilC: Yes my orginal statement should have said when pages are added to the site the home page and/or sub pages loose. I corrected that statement in a later post.

I came up with the same thing in my diagrams as you 2 below examples.
When you use a hierarchically structure the home page gains 1 or 2 real PR for every page you add when you are feeding 1000 real PR into the home page. The sub pages all drop. In the fully meshed site structure (which is the way most smaller sites are) the home page real PR dropps with ever page that is added.

Thanks for adding your comments.

PhilC

04-12-2005, 11:37 PM

Sorry, I didn't spot your corrected statement, Bob.

I agree - a fully "meshed" (all-to-all) link structure does produce losses on all existing pages for every new page that is added.

An all-to-all link structure can be likened to a family who sit down to dinner. There are 5 family members, and the dinner consists of potatoes and nothing else. There are 15 potatoes in the bowl - 3 potatoes for each person (they each brought 1 potato with them and the rest were provided by "meals on wheels", of course :) )

But a visitor arrives and sits down to eat with the family. The visitor brings 1 potato and puts it into the bowl. Now there are 16 potatoes to divide equally amongst the 6 people at the table. The result is that each of the 5 original family members gets less to eat than s/he would have had if the extra person hadn't arrived, even though there are more potatoes in the bowl.

If enough extra people arrive, each bringing only 1 potato, then the Toolbar potatoes for the original family members can even be decreased. :eek:

The formula for the calculation is:-

1 potato 2 potato 3 potato 4, 5 potato 6 potato 7 potato more

where 'more' is set at 0.85, and is the dampening factor (often called gravy) that prevents the potatoes from drying up and up and up.

bobmutch

04-12-2005, 11:57 PM

JohnW: >>>What benifits do you see with having higher PR on the home page as far as SEO goes<<< "I think we can all agree that getting good links from good pages enhances a pages “ability” (lets call this the authority score or PR) to rank for relevant keywords, and, to affect the authority score of other pages it links to."
I would agree that when real PR carried more weight in the ranking algo that is was a good thing to have high PR on a page, and as real PR may still carry some ranking weight it is still a good thing. But I think you would agree that if you are targetting 20 or so keywords you are not going to be in most cases targetting them all on the home page. If fact I would say in general 1 main keyword and perhaps 2 or 3 minor keywords targeted per page.

Therefore the idea of channelling real PR to the home page is not the way to do it. You would want to channel you real PR to the pages your have optimized for your keywords.

The idea of pumping the home page up high in real PR so it has a "authority score" and in then in some way that home page is to affected the "authority score" of the other pages linked to it is a concert that I have never read nor heard of.

I would think that real PR should be channelled to all the pages that you have optimized for keywords. Not just to the home page.

If the idea of channelling as much real PR as possible to the home page doesn't make SEO sense for the above reason then the idea that a site structure that bleeds real PR from the home page is faulty falls with it.

"So it seems obvious that there is a authority score of some kind, PR, whatever you want to call it, and that it is good to have more of it and bad to have less of it."
I would say that a site structure that bleeds the real PR off any page and it doesn't go to a page that you have optimized to rank is not good. I would not how ever hold that having a structure that bleeds real PR off the home page to other pages that are optimized to rank for keywords is a bad thing. I would say that is a good structure.

"That’s also why earlier, I said full mesh navigation where every page links to every page, is bad SEO."
If each page within that full mesh is optimized to rank for keywords I would disagree. You have 10 pages and they are all optimized for 2 or 3 keywords each, then you would want the real PR spread around evenly in most cases.

John I just think that the knee jerk reaction that bleeding real PR away from the home page is a bad thing is a faulty reaction. If real PR still has ranking weight it should be channelled to all pages you want to rank with.

Another issue is if you want to pick up the ranking weight of keyword in the URL you usually are not going to be able to do that on the home page. In such a case you are going to want to channel your PR to your this-is-my-may-keyword.html page not to the home page.

bobmutch

04-13-2005, 12:10 AM

Oh PhilC you are just to funny. Through working with the PR forumla I found the best way to channel real PR to the home page and at the same time keep a workable navigation is to have 4 main pages fully meshed. Then run a link from those 4 main subs to all your other pages and have those pages link back to the home page only. Then have most of your links come in on your home page. I have done this with my site and it seems to have worked very well.
http://dir.google.com/Top/Computers/Internet/Web_Design_and_Development/Promotion/Search_Engine_Optimization_Firms/S/

dannysullivan

04-14-2005, 07:53 AM

Let me preface my responses to say that, as always, I'm wary of depending on anything published by Google as proof of exactly how they actually operate. I'm especially wary of relying on the main paper about PageRank that was published years ago. Things have progressed since then. We know they don't count certain types of links. We strongly suspect they weight the PR value of some links depending on the position on a page and repetition as seen on multiple pages (nav links like we have on the left-hand side, looking the same, same anchor text, pointing to the same exact page simply might not be counted for as much).

Next, let me give you the perspective/response I've had when people tell me they are worried about PR "bleed" or losing PageRank by linking to their own pages or to other pages.

I've encountered this directly because at our conferences, Google will do its usual presentation on how classic PageRank calculation works, showing a page that links to five other pages, with each of those other pages getting 1/5th the value of the PageRank score that the page linking to them has. Now make that 25 links, and each link gets less of a share 1/25th.

That's classic PageRank. It says nothing that the page itself that's having all the outbound links is somehow leaking any type of its own PR score. Instead, it's an issue that if you are more selective about what you link to, you are able to better transmit what authority you have to share.

For the search marketer, this is why getting a link on a page with hundreds of links is probably not worthwhile. Each additional link means that all the links have to share less of the PR pie.

Internally, this means you might want to be more selective on how you link to your own pages. Got 1,000 pages? You could have every page carry links to all of them. Do that, and then each page will only pass on 1/1000th of its value to other pages. Be more selective, and you can have a bigger impact.

That's especially important with your home page, which is generally your most important page in the site, carrying the highest PR value. Link to all your pages, and you might be wasting an opportunity to transmit more importance to a few key pages that you are most worried about.

OK, from my view, something to sort of keep in mind but which shouldn't override the dominanting factor. Link to whatever pages you want to link to on your pages, and let what makes sense for your human visitors dictate that. It's bad to have upteen million links on your home page because you aren't giving your human visitor any type of guidance about what's important. Instead, list the stuff that is important -- and don't list everything. As people drill into the site, link to your own content and external content as it makes sense.

So -- does adding pages hurt your PR value? No. Just having more pages doesn't hurt anything, in any way, shape or form that I can see. Search engines don't even know whether you have more pages until they actually find them by following links, which leads to...

What is really meant is whether adding LINKS to your new pages (or other pages) will hurt your PR value. As said, I don't believe that having new links on a page will cause that page to drop in PR value with the key exception of whether Google considers the outbound link to a "bad neighborhood," in which it then might decide to penalize you. Fair to say, I think, this isn't a worry for the very vast majority of people.

Adding links potentially could hurt you in that dumping a ton on a page might cause the existing importance of links to your own pages that previously were there to get diminished/diluted. But then again, external links to those pages might carry a heavier weight. It's impossible to know exactly what might happen. Fair to say, adding a ton of links to any page is probably bad from a human reader perspective, so don't do it at least for that.

Both these examples really only show us that if you throw enough links at your home page you can off set the real PR bleed to a large number of inside pages.
I don't agree, sorry. You're making it sound like it's unusual for a site to have inbound links. Yep, SEW has a lot of them. But with any site, inbound links could offset any type of PR bleed you want to lay claim to. As said, I don't feel the page itself bleeds any PR score just by outbound linking, of course. But let's say this is indeed the case. So you've got bleed when you link to more of your pages -- but the degree of that might be minimal, might be offset by inbound links from external pages, might be severe or might change tomorrow.

I don't know where you are getting the idea what I or anyone has stated linking back to the home page is not a good think to do if the objective is to have high PR on the home page. Common sense tells us that is a good thing to do.
It came from this:

I have yet to find a site structure that has inbound links to the home page where you can add pages in the site with out bleeding real PR from the home and or sub pages.
That pretty much made me think you were saying there was no way to have a site that links back to its home page AND adds new pages AND DOES NOT bleed PR. And as said, I've already pointed out SEW and Amazon as just two examples where I think you have a site structure that allows exactly this and which does not hurt the PR score of the home page. If the argument is that inbound links are what offsets that, OK -- but that means any site might have inbound links that could do a similar offset.

the issue that is being discussed as I see it is that the more pages you add to your site (with no change in the real PR vote to your home page by IBL's) the real PR will be voted and retained by those new pages during a real PR update. This will decrease the real PR on the home page in most cases, and will decrease the real PR on the home page and main sub pages in all cases.
So if all things were frozen in time (which never happens, but I'll roll with it), you add more pages to your site (and correspondingly, more links to those pages from within your site in some type of manner) and thus will cause the internal PR you were totally in control of to be more diluted. OK, maybe that will decrease the PR. Maybe not.

Did you add links to every page? Were the links in navigational areas that might have been ignored? Is Google already ignoring or discounting your internal linking? Did you add new links but perhaps remove some other ones? Were the pages even linking to your home page? All of this and more can have an impact -- and the inbound links are not a factor that freezes in time. They are a factor that potentially far outweights your own internal linkage.

So my worry is that this started with this statement:

The fact of the matter is the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page.
That leaves a newbie thinking "oh no," I'd better not add any new pages -- if I do, that's going to hurt me. I don't think that's a fact at all. Indeed, if you're adding good new pages, those pages might attract good new inbound links, which themselves will transmit importance to your other pages based on linking patterns.

That statement came after this, of course:

Completely impossible. The idea that you can generate toolbar PR from lots of internal links is a complete myth.
And I think that's a point well taken, which puts the earlier statement in better context. I want a high PR score for my home page. Can I just add a bunch of pages and link back to myself to get it? My view is probably not. Those new pages won't have much importance to transmit, if they haven't gained many links to them to begin with. In addition, I strongly feel Google will downplay the value of internal links specifically to prevent people from just whipping up a bunch to get a high PR score. So no, I don't think adding new pages will magically get you a PR spike. But I would add new pages if you've got good content and reason to do them -- and down the line, I'd argue those pages could benefit you overall.

PhilC

04-14-2005, 08:58 AM

I agree that the original PR equation has probably changed since it was published, but I don't agree that it has changed in such a fundamental way as to take account of such things as weighting the PR value of some links. I'm not saying that they don't that; I'm saying that I doubt that they do it because, imo, there are better places to weight certain types of links than in the PR calculations.

Also, remember that Google didn't publish that "main paper about PageRank". It was published by two college kids who had no intention of starting their own search engine at the time. It was part of their college work, so their is no reason to suspect it as being not entirely true. Later they tried to sell their technology, but nobody wanted it, so they started their own engine.

Can I just add a bunch of pages and link back to myself to get it? My view is probably not.Actually, you can. The more pages you have in the site, the more PageRank there is in the site, and the more there is to channel to wherever you want it. But it takes an awful lot of new pages to change the Toolbar PR of the target page. I once added over 20,000 pages to a site, which took the home page from PR4 to PR6. I don't know where the thresholds were because the new pages were genuine and I wasn't looking to see how many were needed. And, of course, I had no idea whether the home page started as a strong PR4 or a weak one.

This discussion is about the technicalities of PageRank, but I think the most important thing that should come out of it is *not* to be too concerned about PageRank. It is probably still a small ranking factor, and it doesn't hurt to take account of it, but there are far better ways of moving up the rankings than chasing every ounce of PageRank. Unfortunately, PageRank is still an unreasonably big thing in many people's minds.

Mel

04-14-2005, 09:25 AM

True PR probably does have some ranking value, however small, and Google still rely on it for many other things like how often a page gets crawled and ordering and replication of the Index shards, but its real power lies in the Urban Legends that are still being perpetuated online and which have knock on economic effects.

Purchased links are priced by PR ( though no one will admit to that publically) and if you take a given web site with a PR8 attached it will fetch a higher sale price than the same site with a PR5.

Personally I think that Googles preactice of PR is just as much about Public Relations as it is about Page Rank.

seomike

04-14-2005, 10:26 AM

I did a mod rewrite for an ecommerce site whos home page had a solid PR 5 and internal pages where completely invisible. After the PR update the home page PR dropped to a 4 all the pages on the root went to a PR 4 and all pages in a sub folder gained a PR 3. The backlinks grew by a few numbers and lost no old links.

As per the rankings they went through the flippin roof even before the PR update happend due to the sites structure. Not only did we retain the same positions for the home page with a PR drop but the internal pages all ranked for brand names in Yahoo MSN and Google something that her old site couldn't dream of doing.

My point is this. Forget about bleeding PR, build the site correctly and you will rank.

bobmutch

04-14-2005, 03:12 PM

dannysullivan:

"I'm especially wary of relying on the main paper about PageRank that was published years ago."
I can understand that. I however don't think that the PR formula has changed to such a degree that PR vote, PR channeling or PR bleed is no longer works the same way.

"Link to whatever pages you want to link to on your pages, and let what makes sense for your human visitors dictate that."
I fully agree! The idea that if your site structure bleeds PR from the home page that there is some thing wrong with that structure is nothing more that silliness in my mind.

"So -- does adding pages hurt your PR value? No."
I agree. Having "hurt" PR is not a phrase I have ever used. Will it bleed real PR from the home page? in many cases yes, certainly in a fully meshed structure. Will it bleed real PR from your main sub-pages? In every case that I have tested yes. Is this a "bad" think? No.

I hold that there is no big reason to want to channel all your PR to the home page other than for bragging rights of a high PR or selling links based on high PR.

"What is really meant is whether adding LINKS to your new pages (or other pages) will hurt your PR value."
I have no idea. It is a concept that is completely foreign to me and introduced into the thread by someone that seems to hold that bleeding PR from the home page to the internal pages is a bad thing.

"You're making it sound like it's unusual for a site to have inbound links."
Well if you think what I am saying promotes that concept you are clearly misunderstanding what I am saying. I re-read may be in order.

Personally I think my point was a clear point, how be it easy to make. Your claim that a site in which the toolbar PR has not dropped, and has had an increase in the number of site pages as support for "adding pages doesn't bleed PR for the home page" while there is been a substantial increase in IBL's is a poor argument in my opinion.

The increased IBL links pointed at your home page votes real PR that page. To say that your toolbar PR is not dropping when you add 1000's of pages to my site while the added IBL's are voting move real PR to your home page proves nothing.

If fact there are so many variables it would be hard to find any single example you could use for sure to prove or disprove where adding pages bleeds real PR or adds real PR.

Even if you tried a test on a site that would have no increase in its IBL's one side could say the toolbar PR dropped at TBPR update time because the real PR vote was less from the IBL links and the other site could say the PR increase only because the real PR vote increased.

What would have to be done is create two 10 page sites that have the very same structure, the very same number of IBL from the same sites, then add 10,000 pages to one of the sites and leave the other site the way it is and wait for a toolbar PR update and see what happens to the toolbar PR on the homepage and main subs. Then where would be no wriggle room.

I don't know where you are getting the idea what I or anyone has stated linking back to the home page is not a good think to do if the objective is to have high PR on the home page. Common sense tells us that is a good thing to do.It came from this:I have yet to find a site structure that has inbound links to the home page where you can add pages in the site with out bleeding real PR from the home and or sub pages. Danny linking back to the home page votes PR to the home page. I could see you misunderstanding what I said and thinking that linking OUT from the home page is a bad thing but really, how can you get that linking TO the home page is a bad thing over a statement where I note that nature of the beast is that real PR is bleed when you link OUT from the home page.

When people want to channel real PR to the home page what they do is minimize the links out on a page and just link back to the home page. This way the vote is 1/1 and not 1/20. So where you came up with the idea that any of my statements could infer that IBL's from inside pages to the home page are not a good thing if the objective is to channel real PR to the home page is beyond me.

"That pretty much made me think you were saying there was no way to have a site that links back to its home page AND adds new pages AND DOES NOT bleed PR."
Not at all! I have stated at least twice in this thread that a hierarchically structured site will add real PR to the home page but bleed it from the main sub pages. My article on my site "Page Rank Explained" shows diagrams, live calculations and code that supports this. In the last thread we had on this issue I posted the diagrams and clearly stated that.

"If the argument is that inbound links are what offsets that, OK -- but that means any site might have inbound links that could do a similar offset." Correct.

"So if all things were frozen in time (which never happens, but I'll roll with it), you add more pages to your site (and correspondingly, more links to those pages from within your site in some type of manner) and thus will cause the internal PR you were totally in control of to be more diluted. OK, maybe that will decrease the PR. Maybe not."
I agree. Just because the real PR drops some doesn't mean the toolbar PR will move. But again the toolbar PR is only important when you are selling links are want it for bragging rights. It is the real PR that has ranking weight (well used it anyway).

But in this I am not saying toobar PR doesn't matter. I hold that what is important is what the toolbar PR indicates you have - IBL's, not the toolbar PR itself. Its just a rough indicator of the value of your IBL's.

"That leaves a newbie thinking "oh no," I'd better not add any new pages -- if I do, that's going to hurt me."
I agree. I need to qualify that statement every time I say it. I have clearly noted several times that voted real PR from the home page is not a bad thing but can be a good thing. If real PR has ranking weight you want it on the pages you want to rank high for. I should also qualify that statement you are quoting. Real PR is voted about from the home page and/or the main sub pages as in a hierarchically site structure the real PR increases just a bit.

External link 1000RRP ---->A 3064RRP B/C 1302RRP
External link 1000RRP ---->A 3065RRP B/C/D 869RRP
External link 1000RRP ---->A 3065RRP B/C/D/E 651RRP
External link 1000RRP ---->A 3072RRP B/C-U 130RRP

The real PR on the home page goes up and the sub-pages drop when you add pages with a 2 way link from the home page.

"And I think that's a point well taken, which puts the earlier statement in better context. I want a high PR score for my home page. Can I just add a bunch of pages and link back to myself to get it? My view is probably not."
Besides a few misunderstandings I think we agree on most of the discussed issues.

Let me sum up my views that we have discussed here.

1. You increase the over all real PR of a site every time a page is added.
2. Your decrease the real PR of many pages every time a page is added to a site.
3. Adding pages with 2 way links to the home page in a hierarchically site structure will increase the real PR of the home page but decrease the real PR of all the other sub pages.
4. Real PR can't be hurt it can just be channeled, voted, and bleed. None of these things are inherently "bad" or "hurt" a site.
5. Adding a page to a site with a fully meshed site structure will always bleed PR from the home page.
6. When real PR had ranking weight, that real PR would be best channeled to the pages you want to get good rankings on even if that meant a bleed of real PR on the home page.
7. The idea that bleeding real PR off the home page is bad or hurts a site is a misconception that has been born on a wave of PR hype, link selling based on PR and the concept that having a high PR site means you are successful.

bobmutch

04-14-2005, 03:32 PM

seomike: "My point is this. Forget about bleeding PR, build the site correctly and you will rank."
I totally agree. With real PR being devalued so much in the last couple of years, to the point many think it has little or no weight left, that your site structure and page to page links shouldn't be dictated by real PR bleed at all. Well unless you what a high PR on your home page for bragging rights : )

bobmutch

04-14-2005, 09:00 PM

PhilC: "Actually, you can. The more pages you have in the site, the more PageRank there is in the site, and the more there is to channel to wherever you want it. But it takes an awful lot of new pages to change the Toolbar PR of the target page. I once added over 20,000 pages to a site, which took the home page from PR4 to PR6. I don't know where the thresholds were because the new pages were genuine and I wasn't looking to see how many were needed. And, of course, I had no idea whether the home page started as a strong PR4 or a weak one."
It would really be hard to say that those 20k of pages took the site from TBPR4 to a TBPR6. You could have had an increase in the IBL's. The current pages that were giving you your IBL's could have been voting more real PR on the update you jumped from PR4 to PR6 than on the last PR update, so it is hard to say that 20,000 of pages did that.

JohnW

04-14-2005, 10:36 PM

>so it is hard to say that 20,000 of pages did that.

True, but we keep seeing it happen. Looks like you guys are on the right track, not that I recommend this as a tactic - I don't. But the principles are good to keep in mind when designing nav structure.

projectphp

04-14-2005, 11:19 PM

We strongly suspect they weight the PR value of some links depending on the position on a page and repetition as seen on multiple pages (nav links like we have on the left-hand side, looking the same, same anchor text, pointing to the same exact page simply might not be counted for as much).
Can I ask why you think that, Danny? would seem to me that PageRank is a massively processor intensive algo to work out already, what with multiple iterations required. Surely adding in additional factors would only slow it down?

I have a related question: many moons ago, there was a paper on speeding up pagrank calculations: http://www.nsf.gov/od/lpa/news/03/pr0356.htm and full paper here http://www.stanford.edu/~sdkamvar/research.html.

Wasn't one of the ideas to calculate PageRank only for sites, and then distribute the PageRank within the site? If that is the case, and the increase in the speed of calculation is pretty significant, how does that change everyone's view?

bobmutch

04-14-2005, 11:34 PM

Mel

04-15-2005, 12:01 AM

JohnW: "True, but we keep seeing it happen."
Content brings rankings, rankings brings traffic, traffic brings IBL's, IBL's vote RPR. I would expect in most cases that increasing content by 5,000+ page is going to increase the TBPR.
I guess what we need to do is get the owners of 5 - PR8 sites to offer 2 links on a PR8 page, put up 2 sites the very same and then add 20,000 of pages to one, point the 5 PR8 pages at each of them and see what happens toolbar PR update day.

LOL, just do the math. If your internal navigation structure is set up in one way adding pages has to increase PR of the home page, if its set up in another it will not.

I really doubt that there is much if any correlation between traffic and IBLs.

PhilC

04-15-2005, 01:04 AM

It would really be hard to say that those 20k of pages took the site from TBPR4 to a TBPR6. You could have had an increase in the IBL's. The current pages that were giving you your IBL's could have been voting more real PR on the update you jumped from PR4 to PR6 than on the last PR update, so it is hard to say that 20,000 of pages did that.It may be impossible to prove it, but when the site sat around with PR4 on the front for several years since the toolbar came out, and suddenly jumped to PR6 following the addition of 20k+ pages, thinking that the new pages caused it is a reasonable assumption to make, imo.

dannysullivan

04-15-2005, 05:24 AM

Bob, on the entire "linking back to the home page" issue, I've I misunderstood what you were saying, my apologies. I already listed the exact quote where I got that impression, but from your last post, sounds like that's not an issue.

If fact there are so many variables it would be hard to find any single example you could use for sure to prove or disprove where adding pages bleeds real PR or adds real PR.
Let's get very precise in our language. Adding PAGES isn't going to impact the PR scores of anything. I have pages on my site Google has never seen, nor that anyone knows to link to. Only links are counted. Put up 20,000 pages, and if there aren't any links to any of those pages anywhere, no impact will occur.

Adding LINKS will have an impact on PR. If you add a bunch of pages to your own site, then LINK to those pages, that has an impact. As for whether that's a positive or negative, I think exactly as you, there are so many variables that it is hard to say.

That's why when you said flat out that adding more pages (IE links) to your site will cause real PR to be "voted away" from your home page, I disagreed. There are so many variables that I don't know how you can say that. Adding more pages might generated new inbound links, which might increase your PR scores. Isolating the site and assuming no inbound links isn't realistic for most people. It's really an issue for a brand new site that thinks they'll add a ton of new pages and generate good PR that way. Without inbound links to help, I don't think they'll be successful.

Having "hurt" PR is not a phrase I have ever used. Will it bleed real PR from the home page? in many cases yes, certainly in a fully meshed structure. Will it bleed real PR from your main sub-pages? In every case that I have tested yes. Is this a "bad" think? No.
So this is probably my week to be sensitive about language :)

Bleed is a pretty charged word. When you say bleed, I hear hurt. It sounds like a bad thing. And my fear addressed above is that a newbie will read that there can be PR bleed, be fearful and decide they should add new pages at all. Personally, I've used transmit sometimes when talking about these issues -- a page can transmit its authority to other pages, but the more pages you're linking to, the less authority it can share among them all. That doesn't make it sound like the page itself is being harmed. My view, of course.

1. You increase the over all real PR of a site every time a page is added.
Disagree. Just adding a page does nothing. Just adding a link to the new page from an existing page or page within your site is impossible to calculate from where I sit as an increase or decrease given we don't know where the link is, how many are added, whether it will be counted by Google or not and the impacts of inbound links. Finally, to my knowledge, a "site" still does not have a PR score with Google. Each individual page within a site has its own score.

2. Your decrease the real PR of many pages every time a page is added to a site.

Disagree again because of all the unknown variables involved.

3. Adding pages with 2 way links to the home page in a hierarchically site structure will increase the real PR of the home page but decrease the real PR of all the other sub pages.

Maybe. I assume you mean this is because since every page in the site will have a link to the home page, removing that might let them share more PR within internal pages. Then again, if some of those internal pages have some new good inbound links that come in, that bumps them up. Linking back to the home page bumps it up. It links back to other pages, and goodness might flow around. Lots of variables still involved to flat out say this.

5. Adding a page to a site with a fully meshed site structure will always bleed PR from the home page.

Don't know the definition of what you mean by fully meshed, so can't really comment. Fully meshed may mean every page links to every other page. Them's a lot of links, if so! But again, all the variables makes me think it's hard to say this.

6. When real PR had ranking weight, that real PR would be best channeled to the pages you want to get good rankings on even if that meant a bleed of real PR on the home page.
If this means linking to other pages hurts the PR score of the home page, I've never bought into that. Never heard Google suggest it, either -- though as said, you can't always believe what they say. If this means linking to other pages means each link reduces the PR score that can be transmitted to them all, I agree.

Can I ask why you think that, Danny? would seem to me that PageRank is a massively processor intensive algo to work out already, what with multiple iterations required. Surely adding in additional factors would only slow it down?
So we know Google doesn't count certain links like guestbooks because I think they've both said it and others have been able to verify it in some cases, which I recall coming up in forum discussions over the years. They don't say exactly what links may or may not be counted, of course. As for Google not counting some links as heavily, this comes from having talked with Google over the years when various changes have happened. You get the usual Googlespeak of "you can imagine that we might want to count certain links with less weight" or "you can imagine we might not count certain links at all" responses. That doesn't mean yes, of course -- hence me saying we suspect.

As for increasing what needs to be computed, I don't think that's much of an issue. Google's got a lot of horsepower. In fact, I still suspect they are computing ranking usual multiple search algorithms on the fly. Aside from that, PR scores themselves I believe most feel are still calculated not in real time (as with Teoma) but instead at set points. A lot of this stuff especially came up during the Florida update. Florida Google Dance Resources (http://searchenginewatch.com/searchday/article.php/3285661) had a lot of reading on that from SEW and beyond.

Unfortunately, PageRank is still an unreasonably big thing in many people's minds.
Edit/Delete Message

-and--

My point is this. Forget about bleeding PR, build the site correctly and you will rank.
I agree entirely. The discussion is interesting, but the worry over PR can drive people inside. When I do my Intro presentation at SES, I have a section on link building with a slide that says it's not about the Google Toolbar or PR. I then start off saying that you want links from pages with high PR, right! I think go through all the other factors you need to consider. Is nofollow used? Are these guestbook links? How many links are on the page? What the actual context of the link? And so on. By the end, hopefully I've made you head hurt a bit, so that my Golden Rules Of Link Building (http://searchenginewatch.com/sereport/article.php/3422121) come as a relief. That link is to a 2002 article where I said forget about the toolbar and focus on these three tips:

Rule 1 - Get links from web pages that are read by the audience you want

Rule 2 - Buy links if visitors that come solely from the link will justify the cost

Rule 3 - Link to sites because you want your visitors to know about them

Those are simplistic. They aren't for everyone. Those aggressive in SEO may decide to ignore them entirely. But I think they are good starting places. Goodness knows the Google Toolbar and the mysteries of PR isn't a place I want those new to SEO to begin with.

PhilC

04-15-2005, 10:42 AM

So this is probably my week to be sensitive about languageYou can say that again ;)

Without going through all the replying to quotes bits, let me just make some generalisations here in response to your post, Danny.

When people talk about adding pages, it is understood that they will be linked into the site, so the differentiation between adding a page and adding a linked page isn't necessary.

Adding a page to the site *does* add PageRank to the site, regardless of new IBLs, changes in the value of IBLs, or any other factors. It is simple mathematics. Perhaps other factors removed more PageRank than the new page added - it doesn't matter. The fact is that adding the page added PageRank to the site.

It is true that, to the best of anyone's knowledge, sites don't have PageRank - pages do. But you are missing the point. We have sites, and each page within the site normally has some PageRank. *We* can think of it as the PageRank within the site, and we can move it around the site more-or-less as we wish. That is what is meant by the PageRank within the site, or the site's Pagerank.

Just as an aside, when people refer to "a PR5 site", often when looking for link exchanges, they (intentionally) misuse the idea of a site's PageRank to make the exchange sound better than it is.

Back on track:- The fact of the matter is that adding a page to a site *does* increase the PageRank within the site, and, if the site already has some additional PageRank from IBLs, then the new page *will* drain some PageRank from the existing pages as a whole. Precisely which pages lose, and which actually gain, depends entirely on the linking structure within the site. It's simple mathematics. Other factors don't come into it. In fact, IBL values/additions, link weights, etc. are just red herrings.

All this is assuming that there have been no radical changes to the original PageRank equation. For reasons that I stated earlier, we can consider the papers that Brin & Page published, at the time they published the main PageRank paper, as being entirely accurate. Any suggested changes to the original PageRank equation/formula are just speculations, and it is unwise to base an understanding on them.

So we know Google doesn't count certain links like guestbooks because I think they've both said it and others have been able to verify it in some cases, which I recall coming up in forum discussions over the years.If that were so, how come they recently instituted the "rel" property if they already had it covered? If it is so, what don't they count guestbooks links for - PageRank calculations or rankings or both? They could drop certain links from the PageRank calculations, but it's speculation to think that they do so, and it's gross speculation to suggest that there are all manner of reasons for dropping links (except "dangling" links) or using modifed link weights for the PageRank calculations.

bobmutch

04-15-2005, 01:00 PM

Mel: Well what I was suggesting was to set up the site structure to channel the real PR from the new pages to the home page. The only structure that I have found that does that is a hierarchically site structure.
With there being a 101k limit on the size of a file you are going to need to put the new 5,000 to 20,000 files in a looped structure to mazimize the real PR vote to the home page.
So you run between 500 and 1000 links on your home page and then run 500 to 1000 loops with each page having a link to the home page.

home-->loop1-->loop2-->loop3-->loop4-->loop5-->loop6-->loop7-->loop8-->loop9-->loop10-->home
Files loop1 through loop10 would all have a link to the home page.

You are going to have to put alot of high TBPR links against the home page to induce the GoogleBot to crawl that.

Again showing that what needs to be done create toolbar PR on the home page by adding pages is counter productive.

bobmutch

04-15-2005, 01:03 PM

PhilC: "thinking that the new pages caused it is a reasonable assumption to make, imo."
Do you remember the structure you used to add them. Could be the the IBL's that you had voted more real PR to the site then the last RPR/TBPR update. I would be interested in knowing the structure of how you added them.

PhilC

04-15-2005, 01:08 PM

No, I don't remember the structure specifically. I released the contents of a database to the engines. People could already get the pages via a form, but engines couldn't. It was pretty much a trail through all the pages, and all the pages linked to the home page as well as to other relevant pages in the trail.

There's no way that IBLs were responsible for the 2-level TPR increase. Yes, I know that a huge coincidence could have occured, but I don't believe in such huge coincidences. It was before the monthly PR updates stopped.

Mel

04-15-2005, 10:59 PM

obiztek

04-16-2005, 01:43 PM

Adding pages does not lower PR if you are not referring to other sites and if your pages are in context and natural with the site subject.

Just adding the dummy pages ( page 1, 2 , ....1000 and so on ) does not helps much.

Google likes natural things and rewards it well

PhilC

04-16-2005, 01:54 PM

I think you've missed what's being said, obiztek.

bobmutch

04-17-2005, 09:20 AM

Mel: "Try this one then Bob homepage-->Pagetwo-->links to 1000 other pages all of which only link back to the home page."
I have tried that site structure before and have found that it reduces the real PR on the home page. I will talk to my math friend and see if he can reduce the PR equation on 1000 pages for me and I will post a diagram with a site with out the 1000 pages and the site with the 1000 pages with a feed of 1000 real PR so you can see what the original PR equation comes up with.

Mel

04-17-2005, 10:11 AM

Then you're doing the math wrong Bob. This is a known standard example and given the conditions the PR of the home page has to be the highest PR on the site.

You can see a detailed and worked out example (http://www.iprcom.com/papers/pagerank/) of this linking scheme here and you will note that the PR of the home page is strongly increased while the value of the 1000 pages is less than 1 percent of the home page PR. There is even a program on this site that you can run to verify it.

This is really simple math, not rocket science, it can be done by anyone with a spreadsheet and an hour or two to spare.

mcanerin

04-17-2005, 12:38 PM

If enough extra people arrive, each bringing only 1 potato, then the Toolbar potatoes for the original family members can even be decreased.

The formula for the calculation is:-

1 potato 2 potato 3 potato 4, 5 potato 6 potato 7 potato more

where 'more' is set at 0.85, and is the dampening factor (often called gravy) that prevents the potatoes from drying up and up and up.

So you are saying I should avoid having a fully mashed site? :D

One note on PR - Some people think that PR is like a big bucket of water and the internet and websites are a complicated mess of pipes and smaller buckets - someone pours a specific amount of water from the top and it filters down to the little buckets underneath, and the more pipes you have to your site, the more water your bucket collects.

The problem with that analogy is that is assumes there is a finite amount of water being poured in by a single source. That's not how it is.

Using this example, every time you add a bucket, it starts out with a little bit of water in it. Not full, just a little bit. But all that water adds up.

Especially if you are a plumber and spend some time thinking about how to organize your pipes.

Just dont block them with crap... ;)

Ian

bobmutch

04-17-2005, 03:23 PM

Mel: Ok you're right. I missed the add a link from each of the 1000 pages to the home page, Sorry.

I presume you are refering to Roger's Example 13. As I have noted a number of times before 331 real PR is nothing to get happy about. Rogers himself has a toolbar PR3 being from 10,000 to 100,000 real PR. So setting up 1000 pages and getting 331 real PR is nothing compared to the real PR of the links you will need to point to the home page to induce Google bot to get all those 1000 pages in to the index and hence into the PR equation.

So I won't call 331 real PR a "strongly increased". Again the concept that you can product toolbar PR on your site by adding 1000's of pages to your site is a myth.

Yes you can add some real PR, the examples on my article "Page Rank Explained" on my site clearly shows that I hold that position. What I have always said it that it amounts to next to nothing in the scheme of things.

You need to point a few toolbar PR5s to get those 1000 pages into the index, and Rogers suggests that the real PR of a PR3 is 10,000 to 100,000. This shows the sillyness of holding the position that you can create toolbar PR on a site by adding 1000's of pages.

Mel

04-17-2005, 10:12 PM

What you have failed to grasp Bob is that the home page in Example 13 started out with a PR of 1.0 and soley by the addition of pages and proper linking that PR was increased 331 times! I call that a strong increase. YMMV

Ok now that we have the math and concept straight, add one incoming link from a TPR5 page (with no other links going out) to the home page of that example and now run the math for example 13 starting with a home page real PR of say 100,000 instead of 1.0 and look at the results. Now you have a home page PR of 3,100,000 which again is a strong increase up to what would probably be a TPR6 soley as a result of one low TPR5 link and 1000 internal pages properly linked. Where did that extra PR come from if not from the internal pages???

bobmutch

04-17-2005, 10:40 PM

Mel: I didn't fail to grasp anything. I am more that aware that every page starts with a real PR of 1. I am also aware that 331 is 331 time more than 1. But what seems to not being grasped here is that 331 real PR is so small it is almost nothing. The PR that you will need to point at the home page to induce the Googlebot to crawl those pages will dwarf your "strong increase" of 331 real PR.

Mel

04-17-2005, 11:03 PM

Looks like you missed the second paragraph mate.

A TPR6 is quite enough to get the homepage crawled and in that example no page is more than two clicks from the home page, enough that all the pages should get crawled.

dannysullivan

04-18-2005, 07:02 AM

When people talk about adding pages, it is understood that they will be linked into the site, so the differentiation between adding a page and adding a linked page isn't necessary.
I deal with a lot of newbies, and believe me, they won't necessarily understand this. I agree with you. Many people will assume that adding pages is the same as adding links. But it's not always so. And more important, adding pages doesn't meant that you'd necessarily add links on all your pages to all the new pages and so on. You might do all sorts of different things in adding links -- which gets back to my earlier point that knowing exactly what would happen is difficult to say.

It is true that, to the best of anyone's knowledge, sites don't have PageRank - pages do. But you are missing the point. We have sites, and each page within the site normally has some PageRank. *We* can think of it as the PageRank within the site, and we can move it around the site more-or-less as we wish. That is what is meant by the PageRank within the site, or the site's Pagerank.
Phil, believe me, I understand what you mean by the site having a sum total of PageRank that can be passed around within it, and I think your language has been very clear on that. But elsewhere, I've seen a site having PageRank written as if the entire site itself has a single score. And many, many newbies make this mistake. They see the site's home page and then assume that's how the entire site is ranked. Then if there's a problem with the home page, they may assume the entire site dropped in PageRank. Just look at what happened with WordPress. The home page went to zero (briefly), but lots of WordPress users started freaking that the entire site had been banned. It hadn't -- it was just that single page. Internal pages at WordPress did just fine.

Just as an aside, when people refer to "a PR5 site", often when looking for link exchanges, they (intentionally) misuse the idea of a site's PageRank to make the exchange sound better than it is.
Agreed, and all the more reason I'm sensitive about this. If people are going to buy links on the basis of a PR score, then they need to understand clearly that the "site" has no PR score -- each page within the site will have its own unique score.

They could drop certain links from the PageRank calculations, but it's speculation to think that they do so, and it's gross speculation to suggest that there are all manner of reasons for dropping links (except "dangling" links) or using modifed link weights for the PageRank calculations.
I tried to qualify all that as well as I could, but I'll say it again. Yes, it's purely speculation that Google might not credit some links as heavily as other links, depending on exactly where they might be placed. I did say that Google won't confirm this. Instead, they do the usual song-and-dance about "you can envision that we might consider such things." That song-and-dance is about as close to confirmation as we generally get. Personally, I'm sure they do such things. Exactly what they count and how they do it, no way to know. I point it out mainly because that spectre is another reason why it's difficult to know precisely what might be happening with links to create PageRank.

As for them not counting certain types of links at all, I am virtually certain that they have said this. But I'll happily through it into the supposition area as well, until I can find time to dig that out somewhere. Of course, nofollow definitely means they don't count some links. But then why even bother with that, if they are already so smart as to discount links?

First, nice PR to get the bloggers off their back. That's the chief reason. Second, they aren't that smart :) It gives them another tool to potentialy fall back on.

PhilC

04-18-2005, 08:54 AM

That's fair enough about making things very clear for newbies. It's a necessity that's sometimes easy to forget.

bobmutch

04-18-2005, 01:23 PM

Mel: "A TPR6 is quite enough to get the homepage crawled and in that example no page is more than two clicks from the home page, enough that all the pages should get crawled."
I think a couple of TBPR5 would do it but a TBPR6 is fine, so we have no disagreement there.

"Ok now that we have the math and concept straight, add one incoming link from a TPR5 page (with no other links going out) to the home page of that example and now run the math for example 13 starting with a home page real PR of say 100,000 instead of 1.0 and look at the results. Now you have a home page PR of 3,100,000 which again is a strong increase up to what would probably be a TPR6 soley as a result of one low TPR5 link and 1000 internal pages properly linked. Where did that extra PR come from if not from the internal pages???"
I think you are missing an elementary concept of the PR calculation. You are claiming that according to Rogers Model #13 when you point a TBPR6 to the home page, and I am willing to go with your example that the home page gets 100,000 RPR voted to it, some how come up with 310 times the PR on the home page.

Mel either it was late and you were not thinking straight when you made this comment, you are meaning something different then what I am reading, or you really don't understand how the PR calculation works at all.
I accept that the addition of 1000 pages in Ian's example #13 product a RPR on the home page of 310. However some how you have changed this to 310 times the home page and now think you have produced 310 times the transfer of a TBPR6's 100,000 RPR to 3,100,000. I don't even think this view merits the time to be explained away but just in case you really do believe this and its just not a slip of the keyboard I will explain how that is impossible.

A link to Ian Rogers Example 13
http://www.iprcom.com/papers/pagerank/#ex13

Each page votes 85% of its RPR out evenly to all OBL's on each iteration. On the first iteration of your example the home page (Page A) will vote 85% of its RPR to the B Page (85,000.85) and the B Page will vote 85% of that plus its own RPR of each to the other 1000 pages (72.2515725). Each of those 1000 pages will vote to the home page 85% of the voted RPR it receives plus it own 1 (62.263836625).
So you see on the first iteration the home page votes out 85,000 RPR and has 62,263.836625 RPR voted back.

We now have Page A = RPR 77,263.986625, Page B = RPR 12,750.2775, and the Spam 1 - 1000 pages with each a RPR 10.987735875. If you total all these figures you should get 101,002.

Do a few more iterations yourself and you will see the total of the RPR stays the same, it is just voted around. There is no multiplying of the RPR of 100,000 that is voted to the home page by 331. A quick look the PR formula will show you that it doesn't work that way.

Also take the time to add another 1000 pages and work out the numbers. Adding pages don't increase the home page RPR at all for this model, they decrease it. Yes it will increase the over all site RPR by 1 for every page you add but it will decrease the RPR of the home page.

Just think about it. Really what you are doing it adding a page that starts with 1. It retains 15% of the PR voted to it so it only stands to reason to unbiased minds that adding pages will vote RPR from the other pages.

The only model I have been able to get to increase the RPR of the home page by adding pages, and I am talking about a model that has a high TBPR link to that page, is a model which has a pure hierarchically site structure. That means pages hanging off the home page with a 2 way link.

As I have noted before this is one of the errors of Rogers and many others models when they don't do the calculations with a few high TBPR pages pointed at the home page. To make these models so they will work in the real world this is required to induce the Googlebot to crawl that page.

When you start pointing high PR links to the home page you increase the RPR on the home page way past a RPR 1. So then when you add pages that all start with a RPR 1 you bleed or transfer or vote RPR from the other pages to that page. In "most" cases this means that the addition of a page will result in the home page voting more RPR to its OBL's than it gets back from the internal IBL's.

Mel this is the 2nd time we are covering this same material. I have stated basically the same things to you and you are basically repeating the same arguments. I suggest you get out your calculator and see how this all works. Read over my page "Page Rank Explained" and then come back with some examples and numbers using the PR formula and show me where my calculations are wrong.

Produce some real work models where adding pages increases the RPR other than the one that I have pointed out. It appears you have not done that. If you had you wouldn't be making the statements you have.

I found myself that when I started creating models and doing the PR calculations that a number of my own misconceptions quickly appeared.

A discussion like this should more about finding the truth of the matter and not just be crafty and wining an argument. This can be done by roll up our sleeves and producing some models and calculations that we can point to and say - "here look at this, this model/calculation proves this or that."

Note: My article "Page Rank Explained" can be found off the home page of the URL beside my handle (up in the custom user title at the start of each post). It is the 13th link under Articles & Information.

Mel

04-18-2005, 09:09 PM

Your math Bob - after the first iteration, the PR of the home page is still 100,000 from the external link plus that voted back to it by the 1000 pages which by your math is 62,263 for a total of 162,263 etc. The PR of the home page will increase with every iteration until the increase is too small to be of consequence and the PR of all pages lower will also increase.

No wonder you think that adding pages decreases PR. ;)

bobmutch

04-18-2005, 10:49 PM

Mel: "Your math Bob - after the first iteration, the PR of the home page is still 100,000 from the external link plus that voted back to it by the 1000 pages which by your math is 62,263 for a total of 162,263 etc. "
Ok your right my math is off on that example. I checked the formula's on my Page Rank Explained page and they are done the right way. Let me do up the formula on this one and see what I come up with.

Ok I did up that site structure with 1, 10 and 20 extra pages. After 40 iterations the final numbers don't change much. Certianly not the 331 times that you were putting forth.
220280, 220283, 220286.

I will put up a page some time this week and post the code and results for 1000 pages. For now I did up the site structure with 1, 10 and 20 pages. Feel free to check the source on the 3 iterations below and let me know if the formula is correct.

http://www.carsinlondon.com/pagerank/prab1.php A B 1 - iterations
http://www.carsinlondon.com/pagerank/prab1.phps A B 1 - source
http://www.carsinlondon.com/pagerank/prab10.php A B 10 - iterations
http://www.carsinlondon.com/pagerank/prab10.phps A B 10 - source
http://www.carsinlondon.com/pagerank/prab20.php A B 20 - iterations
http://www.carsinlondon.com/pagerank/prab20.phps A B 20 - source

I think we will find the very same thing as my examples on my page "Page Rank Explained" show. That adding extra pages to a site at the best only increases the home page a very very small amount.

It would how ever appear that when ever you add a page and only link back to the home page that the PR of the home page will not decrease.

I am still maintaining that the idea that you can increase your TBPR on a home page by adding pages is a myth.

PhilC

04-19-2005, 04:23 AM

That adding extra pages to a site at the best only increases the home page a very very small amount.

It would how ever appear that when ever you add a page and only link back to the home page that the PR of the home page will not decrease.

I am still maintaining that the idea that you can increase your TBPR on a home page by adding pages is a myth.It was already established that adding and linking new pages into a site increases the PageRank within the site, and by organising the link structure, the additions can be used to increase the PageRank(s) of whichever pages(s) are desired. And we established that the TPR increase on the desired page(s) can be significant. It was also established that, if the existing pages have a reasonable PageRank due to IBLs from other sites, then the new pages will drain PageRank from the original ones, so the linkages should be arranged to channel the PageRank to where it is wanted. It was all done and dusted back on page 1.

Adding a lot of pages (1000 is not very many), all linked in the right way, can move a page's TPR up a notch or two.

Btw, it's "TPR" for Toolbar PageRank, Bob - not "TBPR". "TPR" has been in use for about a year, and it's much smoother ;)

Mel

04-19-2005, 04:44 AM

NO bob you do not get the same results when you add one, ten, or forty pages as when you add 1000 pages. ;)

I have done a quick spreadsheet calc of this simple three level model and find that after 20 iterations the PR of page A is getting pretty stable and approaches 261,000 (261,018.07 to be exact) using a damping factor of 0.85, so the PR increased by 2.61 times but still this is a substantial increase for a very biased model.
What it did not do was to decrease.

bobmutch

04-22-2005, 01:07 AM

Mel: "NO bob you do not get the same results when you add one, ten, or forty pages as when you add 1000 pages."
I never said you do get the same results for 1, 10 or 40 pages as you do 1000 pages. Where do you read that? Also that is 40 interations and 20 pages but hey I am the one that is supposed be to proving the best of us can get number mixed up but thanks for joining in, I don't feel so bad now : )

I will tell you one thing for sure. You don't get 331 times more real PR on the home page by adding 999 pages to the site nor do you get the 2.61 times you claiming. Keep in mind it is the differnet between 1 page and 1000 we are looking for. One page gives a home page RPR of 220280 with the formula I am using. So it is the difference between adding 1 page and 1000 pages that we want to look at.

My figures on the difference between 1 and 10 pages is a Real PR of 3. When I when to 20 pages the difference between is a Real PR of 6.

I will be quite supprised if I get a Real PR change of 40,738 between 1 and 1000 pages but lets see what the calculation shows. I should get around to it this weekend.

"...so the PR increased by 2.61 times but still this is a substantial increase for a very biased model."
The Real PR increase of 2.2 times comes with 1 page. What you want to look at is the differnet between 1 page and 1000 pages. That way we see that advantage of adding the extra 999 pages.

The increase from 1 page to 20 pages gets you 0.000027 times more Real PR. And I have a pretty good idea how much it will increase with 1000 pages but lets wait for the math.

PhilC: I am looking into the math to do a model with 20,000 and 40,000 pages also. I have a friend that is a PR expert over in the UK that has a huge math degree. He has helped me work out formulas before.

"It was already established that adding and linking new pages into a site increases the PageRank within the site."
I agree that was my point #1 to Danny.

"...and by organising the link structure, the additions can be used to increase the PageRank(s) of whichever pages(s) are desired."
No one disagreed with that one either. Both these I have noted in this thread.

"And we established that the TPR increase on the desired page(s) can be significant."
Here we disagree. I have yet to see a model where that has been shown let alone established with a real world example.

I would suggest the only what to "establish" it is to do 2 sites the same site structure with the same number of IBL's from the same sites. Then add 40,000 pages to one of them and wait for PR update day. If you can get a PR5 on one site and a PR7 on the other I will fall over backward.

"It was also established that, if the existing pages have a reasonable PageRank due to IBLs from other sites, then the new pages will drain PageRank from the original ones, so the linkages should be arranged to channel the PageRank to where it is wanted."
Again I agree and I don't see any one disagreeing on this either.

"Adding a lot of pages (1000 is not very many), all linked in the right way, can move a page's TPR up a notch or two."
I am going to agree with the notch part but disagee with the "or two" for now. If you have a TBPR 6 and the RPR is right at the top border I agree, 20 pages added up with a link back to the home page gaves us a RPR increase of 6 on the model Mel and I have been discussing, which could be enough to push you over.

But the idea of increasing your TPR by 2 by adding pages I am very doubtful. But lets see if I can put together a formula for the 1000 and rid us of Mel's 2.61 times claim. Even though it is down for 331 times I still don't believe it.

'Btw, it's "TPR" for Toolbar PageRank, Bob - not "TBPR". "TPR" has been in use for about a year, and it's much smoother.'
TBPR or TPR doesn't really matter to me, but thanks for the tip.

PhilC

04-22-2005, 03:48 AM

And we established that the TPR increase on the desired page(s) can be significant.Here we disagree. I have yet to see a model where that has been shown let alone established with a real world example.You don't need to see a model, Bob - just do it.

I think I mentioned before that I have a site that had a PR4 homepage for several years - since the toolbar came out. I added over 20,000 pages and took the homepage to PR6. That was back when the PR was updated monthly, so there wasn't much of a window for coincidental other factors to suddenly take effect.

But the main reasoning is this. You agree that the PR on an existing page can be increased by adding new pages to the site and linking them correctly. If the PR of the existing page can be increased, then enough new pages can take it over one or more TPR boundaries. You can't have it both ways - either new pages, linked properly, can increase an existing page's PR, or they can't. If they can, then more and more new pages can add more and more PR to the target page, and TPR boundaries must be crossed. It's perfectly simple.

Marcia

04-22-2005, 04:20 AM

Before the original issue gets too clouded and obscured: the original issue why this thread was started by being split off as off_topic from another thread, and what the original point was for splitting this discussion off into yet *another* discussion on the same topic as was covered extensively many weeks ago - this was the thread where this started yet_again

PR6 - How many links? (http://forums.searchenginewatch.com/showthread.php?t=5135)

And this was the reply to that member - which was the comment/post that was split off and what this thread here is originally about - again. It was from this post in that other thread this thread was split off from

http://forums.searchenginewatch.com/showpost.php?p=42855&postcount=3

The fact of the matter is the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page. Other things are pertinent, but have we established and settled this particular issue yet?

bobmutch

04-22-2005, 04:40 AM

PhilC: "You don't need to see a model, Bob - just do it."
I think looking at a model is going to be easier. So I would be willing to start there.

"I think I mentioned before that I have a site that had a PR4 homepage for several years - since the toolbar came out. I added over 20,000 pages and took the homepage to PR6."
So some of the Page you were getting your IBL's from picked up a bunch more links or they voted more PR. We just don't why the TPR when up for sure. We can guess what the reason is but guessing is not enough for me.

"But the main reasoning is this. You agree that the PR on an existing page can be increased by adding new pages to the site and linking them correctly. If the PR of the existing page can be increased, then enough new pages can take it over one or more TPR boundaries."
Yes I agree the RPR goes up, but not my much. Say you add 100,000 or 200,000 pages or even 500,000 pages. I would like to see the model that shows how much RPR that adds to the home page. Then we have to deal with the issue of how do you get those pages into the index. How many high TPR links to you have to point at the page to induce the Googlebot to crawl all those pages?
These are some of the reasons I hold that adding pages to a site to increase the TPR is a myth. Lets say the number is 500,000 pages. How you going to get those into the index. I say the RPR vote to the home page the links you will need to point at your home page to induce the Googlebot to crawl all those pages will far out pass the RPR the added 500,000 pages will vote up to your home page.
Also adding 500,000 pages, if thats what it would take, to your site to pick up 1 or 2 units in the TPR is sillyness. Save yourself some trouble and buy a few links.

"You can't have it both ways - either new pages, linked properly, can increase an existing page's PR, or they can't."
I have never denied that adding pages will nto increase the existings pages' RPR. What I have questions is by how much? Is it enough to move the TPR up? What do you have to do to get all the pages you would need to the index so they count in the PR calculcation?

"If they can, then more and more new pages can add more and more PR to the target page, and TPR boundaries must be crossed. It's perfectly simple."
I agree the TPR boundary at some point would be crossed and if you add 10,000,000 I am guessing that you could cross 2 boundaries perhaps even 3.

So I guess I need to qualify my states a bit more because I have never denied it was impossible.

I still hold that adding 1,000 or 10,000 pages to is not going to move the TPR. We have people going around like claiming that adding 1000 pages will increase the RPR of your home page by 331 times or 2.61 times. Its all just sillyness to me.

I am guessing you can vote 300 RPR out of a 1000 pages, 3000 with 10,000 pages and 12,000 with 40,000 pages. 300 RPR will do you nothing. Even 3000 is nothing unless you are at the very top of a TPR border. And I say 12,000 RPR is nothing either.

These figures are probably out to lunch so don't hold me to them. Let me get my calculations done and lets see what kind of RPR would be transfered to the home page. Then we can pick it up from there.

PhilC

04-22-2005, 04:40 AM

New pages take PageRank away from the existing pages, but not necessarily from the home page. With many or most link structures, the home page actually gains.I don't think anyone disagreed with my response to that statement, Marcia, so I thought it was pretty much settled.

Mel

04-22-2005, 04:46 AM

As far as I am concerned Marcia the question was settled several years ago when Google published the PR formula, and I recently ran a spreadsheet to show that if a Page had a Real PR of 100,000 coning into the page from a single and then linked only to one page which in turn linked to 1000 pages each of which only linked back to the home page that the PR of the home page would end up at 261,000 as a result of the internal linking.

Its about a fifteen minute chore to set this up in Excel and run it for those interested. I can send anyone interested a copy of the spread sheet, which is better than publishing it as then you can see the math behind the calcs.

For those who understand math it should be apparent from the PR formula that pages do not lose any PR by linking, they can only gain. It also follows that if you set your site up right you can channel PR around the site.

I suggest that this thread should be closed, as it is no longer serving any useful purpose.

dannysullivan

04-22-2005, 05:44 AM

Sure -- but before we close, last orders! Any final comments from anyone, get 'em in now.

PhilC

04-22-2005, 06:28 AM

Well, all I can suggest, Bob, is that you keep on trying to find the model if it's that important to you.

The bottom line is that adding and linking new pages into a site does add PR to the site, but if the site already has some decent PR dues to IBLs, the new pages will drain PR from the existing pages, although, with the right link structure, certain pages can gain PR.

So the original statement, that "... the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page" is incorrect, and your more recent statement that "you can increase your TBPR on a home page by adding pages is a myth" is also incorrect.

bobmutch

04-22-2005, 12:56 PM

Mel: " ...and I recently ran a spreadsheet to show that if a Page had a Real PR of 100,000 coning into the page from a single and then linked only to one page which in turn linked to 1000 pages each of which only linked back to the home page that the PR of the home page would end up at 261,000 as a result of the internal linking."
The home page of the site has 220280 with only 1 page added to the site.

Even if we go by your calculations, that adding 999 pages produces a RPR of 261,000 on the home page, that RPR is not gotten by adding 999 pages. By your calculations the additional 999 pages produced a RPR of 40,720. And personally I question those calculations. Please post a link to the spreadsheet.

PhilC: "Well, all I can suggest, Bob, is that you keep on trying to find the model if it's that important to you."
I am not looking for a model. I have worked this all out and as far as I am concerned the idea of adding TPR to your home page by adding additional pages is not practical.

What I mean by not practical is that the RPR of the external links you will be required to point at your site to get the Googlebot to crawl the number of new sites pages you will need to vote enough RPR to the home page to see a two unit move of TPR - the PRP of those external links will far exceed the RPR produced.

"The bottom line is that adding and linking new pages into a site does add PR to the site, but if the site already has some decent PR dues to IBLs, the new pages will drain PR from the existing pages, although, with the right link structure, certain pages can gain PR."
Again I agree. Just we differ if this is practical.

'So the original statement, that "... the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page" is incorrect.'
I agree that statement is incorrect. That is why I corrected that statement. I hold that the more pages you add to your site the more real PR is voted away from your main sub pages in all cases and in some cases, like a fully meshed site, it is voted from your home page.

'…and your more recent statement that "you can increase your TBPR on a home page by adding pages is a myth" is also incorrect.'
I agree that statement is incorrect also. It is a myth that it is practical to adding TPR to your home page by adding pages to your site.

Just in case you think I am switching my position I am not. My "Page Rank Explained" article which I wrote about 5 months ago shows in diagrams that you can increase the home page RPR by adding pages.

dannysullivan: Why close the thread? I would like to post the calculation for 1000 pages and see how much RPR is voted to the home page when 999 pages are added using the original PR formula. I will post the code and the calculation for 40 iterations if you leave it open.

Then I would like to tackle the code for 10k, 20k and 40k pages. I think it will be interesting to see how much RPR is really voted to the home page with larger number of pages. Mel has now dropped from 331 times to 2.61 times increase of RPR on the home page when adding 999 pages. And I think if he is reasonable he will need to admit that the 999 pages only increases the RPR on the home page by 0.40720 times as 1 page gives the home page a RPR of 220280 by my calculation (and I posted by code and the 40 iteration calculation above). Also I personally question the calculation that comes up with RPR of 261,000 on the home page.

bobmutch

04-22-2005, 01:55 PM

I got Word out and cranked out the formula. Adding one page gives a RPR of 220280 for the home page after 40 interations. Adding 999 more pages gives a PPR of the home 220610. That is an increase not of 331 time, nor 2.61 times and not ever 0.40720 times but a "substantial increase", to use Mel's words, of 330 RPR or 0.0015 times.

I will product the formula for 10k, 20k and 40k pages after some one takes a look at my formula posted below to make sure I have the calculation coded right.

http://www.carsinlondon.com/pagerank/prab1.php 1 page calculation 40 interations
http://www.carsinlondon.com/pagerank/prab1.phps 1 page code
http://www.carsinlondon.com/pagerank/prab1000.php 1000 pages calculation 40 interations
http://www.carsinlondon.com/pagerank/prab1000.phps 1000 pages code

PhilC: I still agree that by adding pages the TPR can be moved 1 and even 2 units. But I am holding that it is not practical.

dannysullivan: I will crank out a diagram and a few more calculations over the weekend if you want to leave it open.

Marcia

04-22-2005, 02:13 PM

before we close, last orders! Any final comments from anyone, get 'em in now.Yep!

So the original statement, that "... the more pages you add to your site the more real PR (and hence toolbar PR) is voted away from your home page" is incorrect, and your more recent statement that "you can increase your TBPR on a home page by adding pages is a myth" is also incorrect.That about says it.

bobmutch

04-22-2005, 03:36 PM

Marcia: "That about says it."
If you read the rest of the thread you will see I qualified both of those statements.

PhilC

04-22-2005, 04:13 PM

It is a myth that it is practical to adding TPR to your home page by adding pages to your site.You keep modifying your statements, Bob, and it's hard to keep up with you. Now it's changed to not being practical. I assume you mean that Google won't crawl a large enough number of new pages if there isn't a suitable increase in IBLs, or a PR increase in the existing IBLs, which you seem to equate to 'crawl power'. I wonder where you get that idea from.

At the risk of being boring, I'll repeat that I took a site's home page from PR4 to PR6 with the addition of approximately 23,000 new pages. You can claim that there must have been other factors all you want to, but there weren't any. Also, JohnW said earlier in this thread that "we keep seeing it happen.

Bob, this thread was started with a statement of yours which you now agree was incorrect. You changed to a different statement here on the 4th page, but you now agree that that one was also incorrect, so you've added a little modifier to it - "practical". I honestly don't know where you're trying to go with this discussion. It seems that you keep switching your opinion in the hope that somewhere down the line everyone will agree with something. Sorry if that's a wrong judgement on my part.

bobmutch

04-22-2005, 05:03 PM

PhilC: I don't think my core views have changed any. I have had to clarify my statements, and correct them when I stated them wrong.

"Now it's changed to not being practical. I assume you mean that Google won't crawl a large enough number of new pages if there isn't a suitable increase in IBL’s, or a PR increase in the existing IBL’s, which you seem to equate to 'crawl power'. I wonder where you get that idea from."
How do you think that you are going to get 50,000 of pages crawled. Point a TPR2 link at the home page. I think it is a common opinion that pointing high TPR links at your site will get your pages crawled more often and deeper.

"At the risk of being boring, I'll repeat that I took a site's home page from PR4 to PR6 with the addition of approximately 23,000 new pages. You can claim that there must have been other factors all you want to, but there weren't any."
Content brings traffic, traffic brings links, and links bring higher PR. I think it is a pretty common occurrence when you add pages your TPR will go up. The question is why does it go up.

"Bob, this thread was started with a statement of yours which you now agree was incorrect."
I corrected that statement quite awhile back Phil. Re-read the thread.

'You changed to a different statement here on the 4th page, but you now agree that that one was also incorrect, so you've added a little modifier to it - "practical".'
I agreed that statement taken literally is incorrect. I have maintained all along that you can add RPR to your home page by adding pages to your site. That is not a new statement from me and I think you know that.

I hold, that the idea that you can practically boast up your home page TPR is a myth. I have never stated that is is impossible to add RPR to your home page by adding links.

"I honestly don't know where you're trying to go with this discussion."
Well look at the my code and calculations in my last post you will see where I am going. I have pointed out clearly by example using the original PR formula that the idea that you can product 331 times, 2.61 times or even 0.40720 times the RPR on the home page with by adding 999 pages is a myth. It’s more like 0.0015 times.

"It seems that you keep switching your opinion in the hope that somewhere down the line everyone will agree with something."
I have modified my views on PR in the last 6 months sure. I have found out some things I have held where not correct when put to the test. I have had to change the way I have been stating some things to get the point I have been trying to make across clearer. I don't a problem admitting to that.
But my core views have not changed. I have stated a number of things in this thread incorrectly and changed those statements when questioned closer on the statements. Added some qualifier. That is part of life. But its not the way you are portraying it.

Let me ask you a question. If another 39k of pages is added to the example we have been discussing (Rogers example #13 with 100,000 RPR pointed at the home page ) so that we have now 40k of pages. How much RPR is that going to product on the home page. Give me a square answer on that? I haven’t done the calculation yet but I am guessing 12,000 RPR.

Now in your view what are the issues that need to be addressed to get those pages into the index? What does it take to get 40k of new pages into the index. And the RPR that will be voted to the home page - what is your view on what that will do for the TPR?

Instead of crossing my t's and dotting my i's, how about some diagrams, calculations and code to prove your point.

PhilC

04-22-2005, 06:29 PM

I don't think my core views have changed any. I have had to clarify my statements, and correct them when I stated them wrong.You made a statement, Bob, and then you agreed that it was wrong. Then you made another statement, and subsequently agreed that that one was also wrong. Then you added a bit to the second one so that it changed the meaning. I think it's fair to say that you keep changing your views. If you say that you simply keep writing your statement down wrongly, I suppose we may accept it, but we can only understand what you mean by what you write.

How do you think that you are going to get 50,000 of pages crawled. Point a TPR2 link at the home page. I think it is a common opinion that pointing high TPR links at your site will get your pages crawled more often and deeper.Higher PR IBLs aren't a known factor in the crawling of sites. They merely pass PR to the page(s) that they link to. Higher PR pages are a known factor in the crawling frequency within a site, but are they known to be a factor in determining how many pages within a site that Google will crawl?

In answer to your question, just add the pages and they will be crawled sooner or later. It doesn't take high PR for them to be crawled - just a modest PR4 or PR5 will do. As I said, around 23,000 new pages were crawled very quickly on my site with a PR4 home page, and I've had the best part of 50,000 crawled on another site that had PR5 on its home page at the time, although that one did need a little nudge by re-arranging the way they were linked into the site. Just get a modest PR, add the pages, and nudge as necessary.

Content brings traffic, traffic brings links, and links bring higher PR. I think it is a pretty common occurrence when you add pages your TPR will go up. The question is why does it go up.No, Bob, you are mistaken. Content does not bring traffic. Content that can be found brings traffic. And traffic does not bring links. - not in a suitable time-scale, anyway. Certain types of sites are link magnets, but others are not. I think you are trying to suggest that my PR4 to PR6 was due to content, traffic and links, but I've told you that it wasn't.

Well look at the my code and calculations in my last post you will see where I am going. I have pointed out clearly by example using the original PR formula that the idea that you can product 331 times, 2.61 times or even 0.40720 times the RPR on the home page with by adding 999 pages is a myth. It’s more like 0.0015 times.Alright, but who wants to know that particular piece of information? Add 998 new pages instead, and you'll probably find that everything works ok ;)

Let me ask you a question. If another 39k of pages is added to the example we have been discussing (Rogers example #13 with 100,000 RPR pointed at the home page ) so that we have now 40k of pages. How much RPR is that going to product on the home page. Give me a square answer on that? I haven’t done the calculation yet but I am guessing 12,000 RPR.I've no idea, and I've no intention of finding out. I have no interest in following your examples. I'm content with the practicality of experience.

Now in your view what are the issues that need to be addressed to get those pages into the index? What does it take to get 40k of new pages into the index. And the RPR that will be voted to the home page - what is your view on what that will do for the TPR?

Instead of crossing my t's and dotting my i's, how about some diagrams, calculations and code to prove your point.I don't have a point. I only came in to say that your statements were incorrect, and you've agreed with me. Now you don't think that Google will crawl all the new pages, and that's nothing to do with calculations. But instead of doing all the calculations, and working out what may or may not be, just add the pages.

I have a question for you, Bob. What would happen to the final values after the calculations if each new page was not given a starting value of 1, but was given a starting value of 100 instead?

bobmutch

04-22-2005, 11:56 PM

PhilC: "I have a question for you, Bob. What would happen to the final values after the calculations if each new page was not given a starting value of 1, but was given a starting value of 100 instead?"
It is my understanding it doesn't matter where you start the orginal RPR of a page. You will end up with the same numbers in the end if you run enough iterations.
http://www.carsinlondon.com/pagerank/prab1000a.php

You can compare this calucation to the prab1000.php and you will see the end results is the same. Ian, Chris and Markus and you each cover this point in your PR pages. How come you are asking me that?

"I don't have a point. I only came in to say that your statements were incorrect, and you've agreed with me. Now you don't think that Google will crawl all the new pages, and that's nothing to do with calculations."
If the pages doing get crawled that has every thing to do with the calculation as the pages are not part of the calculation until they are in the index.

"But instead of doing all the calculations, and working out what may or may not be, just add the pages."
Because voting 360 RPR to your home page per 1000 pages is sillyness. I would rather just rent a PR7/10 link for $50 a month and vote ~100 times that to my home page. Otherwise if you don't need the 100k of pages it is a complete waste of time.

"I have no interest in following your examples."
Yes I noticed that.

"Add 998 new pages instead, and you'll probably find that everything works ok."
No difference. The same silly ~360 RPR voted to the home page.

"No, Bob, you are mistaken. Content does not bring traffic. Content that can be found brings traffic."
No mistake there at all Phil. You are now just talking circles. What kind of content did you think I was talking about. Your now doing the same thing to me that you corrected danny for doing to you. Any kind of common since would tell you I am refering to content that is found.

"I think you are trying to suggest that my PR4 to PR6 was due to content, traffic and links, but I've told you that it wasn't."
I am clearly and correctly noting that it may have been increase of links or even increase of vote RPR on the same links. But I would have to say like you should be saying - I don't know. You didn't keep the record of the links so neither of us know. I just don't see 12,000 RPR moving the TPR 2 units thats all.

"...but are they known to be a factor in determining how many pages within a site that Google will crawl?"
Lots of high TPR IBL's? Yes there is a good number of people that I have talked to that hold that view.

"You made a statement, Bob, and then you agreed that it was wrong. Then you made another statement, and subsequently agreed that that one was also wrong. Then you added a bit to the second one so that it changed the meaning. I think it's fair to say that you keep changing your views."
My views have not changed in the core. One thing I have had to change my view on is no matter how you add pages to a site, if you ONLY give them a link back to the home page they will not bleed RPR to from the home page. I have not always held that view, or should I say didn't alway relize that to its fullest.

And I will have to admit my view has changed a bit during this thread concerning the amount of RPR that is voted to the home page by added pages. I confess I was in error and thought that there was more RPR voted to pages that my calculations now show me. I have learned that it really is quite a bit less amount of RPR and that I first thought and I can see clearer the sillyness of holding that you can increase the TPR bar 2 units by adding 40k of pages ACCORDING to the old PR formula. Now the real world may work different. One can't really know till they do a controlled study.

One question for you Phil, how much RPR do you think 40k of new pages will product? Can your PR calculature tell you that, or can you come up with a rough figure? And one more, in the scheme of RPR what is you view on 12,000 RPR? Is it a big amount or just a small amound that really won't move your TPR bar?

Mel: You claim your spread sheet shows a increase of 40,720 RPR between adding 1 page and adding an additional 999 pages. My calculation shows only 360 RPR. I have posted my code. Would you point out where my calculation code is wrong please, or at least post your calculation so I can see where I when wrong.

Mel

04-23-2005, 12:33 AM

The question is whether your statement to the effect that adding pages to any site automatically decreases the PR of the home page is correct or not.

The home page of the site has 220280 with only 1 page added to the site.

And I take it that since the home page started out with 100000.85 PR, you now agree that there is a substantial increase by adding only one page to the site and thus you have shown that adding pages to a site does not always decrease the PR of the home page.

Since the premise has been shown as incorrect further discussion seems pointless.

PhilC

04-23-2005, 06:41 AM

Ian, Chris and Markus and you each cover this point in your PR pages. How come you are asking me that?Sorry - I was just finding out how much you know :rolleyes:

If the pages doing [don't] get crawled that has every thing to do with the calculation as the pages are not part of the calculation until they are in the index.Yes, Bob, but why all the stuff about 999 pages and the rest if the only problem is getting the new pages crawled?

"But instead of doing all the calculations, and working out what may or may not be, just add the pages."Because voting 360 RPR to your home page per 1000 pages is sillyness. I would rather just rent a PR7/10 link for $50 a month and vote ~100 times that to my home page. Otherwise if you don't need the 100k of pages it is a complete waste of time.So what's the point of your statements and this thread? Nobody suggested that you *have* to add a stack of pages to improve PR. If you have a stack of pages that can be added (many sites do when they are locked inside a database and behind forms), add them. If there are enough, and if they are linked into the site properly, the TPR of the target pages will improve. If you haven't got any pages to add, get links. It's common knowledge.

"Add 998 new pages instead, and you'll probably find that everything works ok."No difference. The same silly ~360 RPR voted to the home page.It was a joke.

"No, Bob, you are mistaken. Content does not bring traffic. Content that can be found brings traffic."No mistake there at all Phil. You are now just talking circles. What kind of content did you think I was talking about. Your now doing the same thing to me that you corrected danny for doing to you. Any kind of common since would tell you I am refering to content that is found.Yes I am. If it was good enough for Danny, it's good enough for me ;)

"I think you are trying to suggest that my PR4 to PR6 was due to content, traffic and links, but I've told you that it wasn't."I am clearly and correctly noting that it may have been increase of links or even increase of vote RPR on the same links. But I would have to say like you should be saying - I don't know. You didn't keep the record of the links so neither of us know. I just don't see 12,000 RPR moving the TPR 2 units thats all.What can I say? You don't like real evidence, so there's nothing much I can say.

"...but are they known to be a factor in determining how many pages within a site that Google will crawl?"Lots of high TPR IBL's? Yes there is a good number of people that I have talked to that hold that view.I see. There are many people who hold all sort of views that are not based on anything other than imagination. But my question was, are they known to be a factor in determining how many pages within a site that Google will crawl? I've bolded the relevant word this time.

You see, I had a site with a PR4 home page, and it had been PR4 for years. Then, at one particular point in time I added ~23,000 new pages, and Google swallowed them all in very quick time. Now, your are saying that (a) PR4 is good enough for Google to crawl tens of thousands of new pages, in which case, get to PR4 (it's very easy) and then add the tens of thousands of new pages, or (b) that a very big coincidence occured and the home page's PR was suddenly increased at that particular point in time, so that it was sufficient for Google to crawl those ~23,000 pages. But the home page's TPR did not increase, and the ~23,000 new pages were crawled very quickly. It was only after the crawling of those pages that the home page's TPR increased.

I can see clearer the sillyness of holding that you can increase the TPR bar 2 units by adding 40k of pages ACCORDING to the old PR formula.What can I tell you. My evidence isn't good enough for you, but it doesn't matter anyway since TPR is irrelevant. Even so, perhaps you would like to tell us the PageRank numbers where the TPR boundaries occur, so that we have your numbers and the boundary numbers, and we can judge your assertion that jumping 2 TPR boundaries isn't possible.

One question for you Phil, how much RPR do you think 40k of new pages will product? Can your PR calculature tell you that, or can you come up with a rough figure? And one more, in the scheme of RPR what is you view on 12,000 RPR? Is it a big amount or just a small amound that really won't move your TPR bar?Yes, my calculator can do that but not on the web page. But I'm sorry, Bob. I've no intention of getting into actual calculations because it really doesn't matter whether a page can be pushed up 1 TPR or 2 TPRs or 0 TPRs. As long as it's actual PageRank can be increased, TPR is irrelevant.

I still don't know where you're trying to go with this. You've agreed that your statements were wrong, and that the only remaining thing is whether or not many thousands of new pages will be crawled. If you can't get them crawled, get links. It's simple. Why all the calculations?

Added:
Adding thousands of new pages should never be done for the purpose of improving PageRank - TPR or real PR. The purpose should be to target thousands of lesser searchterms, and to get traffic from them. In the year following the addition of those ~23,000 pages, that site earned me ~$45,000 without me lifting another finger - and it's just a sideline.