Interesting results for searches using domain names

...and what they mean

[geoprofi]

Hi all,
with my first post in this very nice forum I've been watching for some time I would like to short introduce myself and to turn your attention to a very interesting fact I noticed while constructing unusuall queries for google (at it seems interesting for me).

I am young SEO/web developer/inet marketing specialist form Bulgaria. I have 4 years of experience in web design and development and around a year and a half experience in the areas of SEO and web marketing.

Well, to the interesting part now...

[edited out][search for a site]

Google returns around 3,350,000,000 results. Well, [Website A] has only one page indexed, so it is not afecting the number above in any way. To be sure that there is no mistake I checked with this one [Google search for website B] - the result is the same. The second site also has only one page indexed.

Well, this query basically states - give me all results you have in your database, but those from site A. And it shows 3,350,000,000 results, while google.com claims they have "©2005 Google - Searching 8,168,684,336 web pages". How could this be? What is happening here?

Let's try this query, to check the results once again, in a bit different manner this time - [Google search website C]
This time we have around 3,320,000,000 which is 10,000,000 results less, which means that site:www.[site C].com should return around 10,000,000 results, right? To my dissapointment it returns the miserable 635,000 results which is far from it.

The question here is how exactly does google treat the "site" command and why it is so unaccurate.

There is another far more interesting point here, though.

Let's go back to the first query - [Google search website A] Now take a look at the the results. What we see is the results, ordered by importance! There a lot of .gov and .org sites, also many big .com sites like careerbiulder.com, yahoo.com, google.com, blogger.com, mac.com. .edu sites are also creddited higher, but only after .gov and .org sites, or at least it seems so to me.

What do you think about these queries, the numbers they show and the results they show?

P.S. Tried the same at yahoo.com, it doesn't work. The guys there must have figured out it is a danger for their algos. MSN.com is a different story. It works like... hmm, I don't know the phrase in English, would appreciate some help :)

It shows they have approx 5,317,569,033 sites in their database. Also it shows that msn gives much less creddibility to .gov sites than google, and even lesser to .edu and .org.
Would like to hear about your thoughts on this one, too

[edited by: msgraph at 7:47 pm (utc) on Aug. 13, 2005]
[edit reason] no specifics pleast [/edit]

[Sanenet]

One thing is web pages (the different content pages within a website), and another is web sites (the collection of content for a URI).

You'll notice that you're only getting the highest ranked page for each site.

[geoprofi]

Hmm, I have to disagree. The results you see are for sure one (max two) per subdomain, the most relevant for the query submitted, but the numbers show all the pages found by the query. If you go to the last page of results you will definately see "In order to show you the most relevant results, we have omitted some entries very similar to the X already displayed.
If you like, you can repeat the search with the omitted results included." And after you click it you will see much more results than that.