23 replies to this topic

[Jim_Westergren]

Hi,

I have just finished some major research and written up an article of a new discovery.

I usually don't drop links in forums but I want the SEO community to know about this, it's exciting:

Here is the article about Google RK values with findings and evidence.

I want some feedback on my findings for my research, thanks.

Edited by Jim_Westergren, 24 February 2006 - 01:22 PM.

[Juliette]

I really hate for my first post to be a negative one , but i tried to go to the page that had your writing on it and it said that "http://www.jimwestergren.com" couldn't be found. I clicked on the link that you provided me with. I'm really interested in reading what you have written so please let me know if i'm the only person with this problem and it's something that i need to fix
Thanks
Jules

[DonnaFontenot]

yep, bad link

[Jim_Westergren]

****.

My stupid webhost changed server and the DNS is still propagating. I can reach it but not the US people ...

I made a temporary copy here:

http://www.seo-fusio...porary-link.htm

Header looks weird but the article kept the format well.

[randfish]

Jim - I just finished reading it and it seems very interesting to me. You're suggesting, basically, that Google's XML feed of search results issues an RK value that should be interpreted as a more accurate PageRank number - correct?

I'd be very interested to see it in use in a tool that pulls from the XML and compare it against numbers we're familiar with on popular sites.

[eKstreme]

This news was broken at WMW, and basically the story is this: Someone noticed that if you get the XML results from the iwebtools PR prediction tool, you see one field called RK. This was hypothesized to be the PageRank and then someone noticed that the RK field was described in the Google Enterprise search help files as "an intenger between 0-10". Then people noted that it correlated fairly well to the known page ranks. This is all explained by Jim.

The upshot is that the RK *could* be the real current PageRank, but as ever, we don't have real proof. Only Google really knows the PageRank. Further, we actually don't know what the field means, whether it is really the PR or not. Google, as ever, is not helping!

As for my theory: it's a good guess but there is no proof yet. We discovered RK tag during a PR update, and the data is constantly: That's not a good experiment! I would be happier if the RK results are checked during times of PR stability, and over a long time period, using lots of URLs.

The other question we haven't answered: The checksum. Many moons ago, when the PR checksum was cracked, we got word that Google changed the checksum. If so, how come the old checksums are still working to show internal PR values?

I'm trying to be negative, but I am trying to be cautious

[Jim_Westergren]

Jim - I just finished reading it and it seems very interesting to me. You're suggesting, basically, that Google's XML feed of search results issues an RK value that should be interpreted as a more accurate PageRank number - correct?

Yes, that is correct.

Based on my findings, for me, it seem very much like PR and there is a strong possibility that in fact this is the internal number Google uses in the SERP ranking.

It is updated regularly and if you for example do a site: search you see that subpages that are more linked to has a higher RK value.

Here is a site: search of seomoz.org using a non-BigDaddy datacenter:

http://66.102.9.99/s...Awww.seomoz.org

Home page RK 6
Next comes the beginners.php with RK 5 and some other sub pages. And then rebecca.php RK 4.

Here you have the tool to see all the DCs with toolbar PR and RKs next to it:

http://livepr.raketf...=www.seomoz.org

The ones showing 7 are the DCs that has the BigDaddy infrastructure imported. There are usually a higher value for those and it might have to do with the new infrastructure and the fact that the BigDaddy datacenters have twice as big indexes (9 billion to 18-25 billion).

The toolbar PR being exported now is from the first week of february which might also explain some differences in RK and TBPR.

I'd be very interested to see it in use in a tool that pulls from the XML and compare it against numbers we're familiar with on popular sites.

Yes, defenitely. Any ideas for specifics? I don't have programming abilities but my partners have.

I also want to set up some kind of RK tracking that measures link popularity, change in SERP and RK values on a list of URLs over time so we can get some more proof of this.

Edited by Jim_Westergren, 24 February 2006 - 12:22 PM.

[kensplace]

The other question we haven't answered: The checksum. Many moons ago, when the PR checksum was cracked, we got word that Google changed the checksum. If so, how come the old checksums are still working to show internal PR values?

I would assume the google checksum for the toolbar PR would be pretty easy to crack, as google offers a toolbar for firefox which queries the PR. As firefox extensions are easy as pie to decompile, all thats stopping someone is the ethics/legal issue of ignoring the google terms&conditions (ie no decompiling, reverse engineering etc etc)

Personally, if I worked at google, I would probably use my 20 percent free time to make up some obsure randomish value, and have it included in some xml results from google, then watch all the seo's jump through hoops trying to beat the system.....


Who knows for sure what the value means? Google does.... We dont, we can only speculate, it could be pr, or some form of ranking result, or something else. It could be live, delayed or made up for all we know. And by the time enough evidence has been accumulated (ie previous predictions come true, or not) google could have changed things again.....


Would google have a reason to include a live internal pr value in any publicly available results? I cant think of one.....
Internal, actual PR only needs to be known within google, there is no reason to make it available outside of google, unless they want to. If they wanted to, then why would they deliberately make toolbar PR delayed?


It could be a red herring, but then again, it may not be, maybe they made a mistake and let something slip out that they didnt mean to. But if thats the case, considering the thread has been up for a while, and they must know about it, why is the result still being shown, unless they *want* people to see that info? (for whatever reason, publicity for google maybe one reason)

Its certainly interesting whatever the case

[bwelford]

I notice, Jim, that the Google definition runs as follows:
<rk> - Provides a general rating of the relevance of the search result

PageRank is a general measure of the "authority" of a given web page, based on back links. It is a single number attaching to the web page.

The above definition might suggest that rk is the relevance of that particular result web page in this particular search. So presumably for a different set of keywords in the search that particular web page might have a different relevance and therefore a different rk value.

Have you been able to show that a given web page always returns the same rk value, whatever the particular key word search? This would then support the notion that it is the same as the single PageRank value that attaches to the web page.

If on the other hand, the same web page can get different rk values in different keyword searches, then rk would be exactly what Google says it is and no more than that.

[Jim_Westergren]

bwelford,

For any query you use, you get the same RK number, it is static.

And about that definition I have written a whole section about that. For me it is clear there was a mistake and misunderstanding on the person that was revising the document.

Here is what I wrote:

Google definition of RK

Let’s look at the definition of <rk> from Google.

From their official “Google  XML Reference“:

“Provides a general rating of the relevance of the search result”

Where does this come from? Seem total wacko, and yes it is a mistake.

I found a very interesting document from Google called: “Google’s Search Results Protocols“, hosted by some guy that mirrors controversial and important documents “that is in danger of censorship”.

And there it says:

Definition of RK:
“Google’s rating of how good a single search result is”

But check this:

In that same document it defines what is a “single search result”.

And it says:

“R - A single search result - Contains a U; an optional T; an RK; any number of F’s; an optional S; and a HAS”

That is the SERP XML!

Every SERP listing in the XML starts with an <R>.

The old definition of R as per that same docuement is:

“A single search result”

The new definition from Google XML Tag Definitions is:

“Provides encapsulation for the details of an individual search result”

So the guy that wrote the new version of this document now called “Google XML Reference”, earlier called “Google’s Search Results Protocols” translated RK:

From:

“Google’s rating of how good a single search result is”

To:

“Provides a general rating of the relevance of the search result”

Which is total wrong, the person didn’t see there was a special definition for “single search result”.

And this has caused headaches for SEOs ever since …

A “single search result” is meant to be a listing in the SERP.

Which means that RK is:

“Google’s rating of how good a listing in the SERP is”

Which is: PageRank!

To further prove the point:

Old version:

    U - The URL of a single search result
    T - The title of a single search result
    RK - Google’s rating of how good a single search result is

New version:

    U - The URL of the search result
    T - The title of the search result
    RK - Provides a general rating of the relevance of the search result

The RK is a static value and has nothing to do with relevance, check yourself.

[Mano70]

Interesting.

I have been running some tests on different websites and pages now, one of those got some strange PR values in this update (the main page holds PR5, subpages PR2). Using Live PR the main page has PR3, and newly indexed pages got PR2 with Live PR which seems more correct. According to my short test your theory may hold water, but I may be wrong...

[DonnaFontenot]

My limited testing also shows the theory has validity.

[praveen]

Interesting find Jim

even though it would take a day or 2 to fully understand it

[Jim_Westergren]

I would have to leave this for 3-4 days in order to take care of clients. But after that I will do some serious observations and testing.

I would appreciate observation/testing data regarding RK from those interested sent to me via PM, as comment on the blog or just posted here. In the data you use it is important to also specify the IP of the datacenter you use as there is a great difference in the RK calculations of BigDaddy and non-BigDaddy DCs.

Thanks a lot and I'll come back with more data later (3-4 days).

[travis]

What sort of value is it ?

If you think of the original PageRank algorithm, this value is most likely a logarithmic value to some base level.

Did you get a large number, any decimal points, integer. What is it ?

[Jim_Westergren]

Google has now put all RK values to zero for all URLs.

Either some temporary glitch OR Google didn't like that value to be public ...

Unfortunately I still have an overload of clients and so I did not manage to do the testings but I am 90% sure it was the internal live PR value based on the observations and discussions I have done. I actually got used to the Live PR FF plugin in my work, well well.

"What sort of value is it ?"

The value was 1-10, just like PR.

[Jim_Westergren]

Matt Cutts read the article and also posted a funny comment on this post

I was very curious and so of course I sent him an e-mail.

His answer:

I’m sorry, I can’t shed light on that at this time.

Best wishes though,
Matt

Well well, seem like Google didn't like that value to be public.

[Aaron Pratt]

Very interesting stuff but now you got Matt Cutts watching you!

I do not know how I feel about Mr. Cutts anymore, I think his spam obsession and sleeping with the enemy is starting to bother me.

:ph34r:

[JohnMu]

I do not know how I feel about Mr. Cutts anymore, I think his spam obsession and sleeping with the enemy is starting to bother me.

I second that ...


PS Anyone else having trouble accessing Jim's site? I can get the main page but anything else comes up blank...

[Jim_Westergren]

PS Anyone else having trouble accessing Jim's site? I can get the main page but anything else comes up blank...

I am sick and tired of that host now and now I am moving. Hope there will be no downtime on the moving part.