Ajax Search API 'estimatedResultCount' wrong answer

[AndreasKruegerDe]

There seem to be inconsistencies between the WebSite results,
and the Ajax results, concerning the __EstimatedResultCount__

Depending on the question,
the (manual) www.google.com search results' 'estimatedResultCount'
are up to about 10 times higher than the Ajax
'estimatedResultCount' !!!

And it is not even always higher, sometimes they are smaller!!!


Please have a look at my measurements:
http://www.andreaskrueger.de/google/Google_EstimatedResultCount_WebsiteVsAjax.zip
http://www.andreaskrueger.de/google/

I simply used the questions 0 1 2 3 4 5 6 7 8 9 10 11 12 13

Here, the difference between manual and ajax 'estimatedResultCount'
is between 91 and 218% (top left diagram).

Sometimes the Ajax answer is even wrong by a factor of 1,000,000
(see the result for "3")

The website result gives the much smoother curve for the numbers
(bottom diagrams - compare left and right), so the website result
is probably the "correct" one, and the Ajax results are faulty!!!



Will you fix that inconsistency?
How long will it take? Can you please send me an eMail when ready,
thanks!
If you will not fix that inconsistency, can you please create an
additional result field that gives some kind of error range?
Something like: "estimatedResultCount=1000,
estimatedResultCountErrorRange=500"

At the moment I could not create the program I am dreaming of! Please
comply.

Many thanks!
Andreas


P.S.: By the way, for these questions 0 1 2 3 4 5 6 7 8 9 10 11 12
13,
there seem to be 9-27 times more "web" results than "images" (top
right diagram).
Quite interesting.

[AndreasKruegerDe]

No answer at all to my question?

Is the wrong number of estimatedResultCount by AjaxApi perhaps a
sinister Google strategy to make programs like mine work less
effective?

As a workaround,
I could quite as well send a normal website-google-request,
and then parse the results page for that number,
only that is more work than necessary, I find.

Please say something.

Please have a look at my measurements:
http://www.andreaskrueger.de/google/

Thanks a lot!
Andreas

On Jul 11, 2:12 am, AndreasKruegerDe <gro...@andreaskrueger.de> wrote:
> There seem to be inconsistencies between the WebSite results,
> and the Ajax results, concerning the __EstimatedResultCount__
>
> Depending on the question,

> the (manual)www.google.comsearch results' 'estimatedResultCount'

> are up to about 10 times higher than the Ajax
> 'estimatedResultCount' !!!
>
> And it is not even always higher, sometimes they are smaller!!!
>

> Please have a look at my measurements:http://www.andreaskrueger.de/google/Google_EstimatedResultCount_Websi...http://www.andreaskrueger.de/google/

[bram...@gmail.com]

I agree, this is very odd and it obviously undermines the usefullness
of the tool.

WHY ?

On Jul 14, 6:34 pm, AndreasKruegerDe <gro...@andreaskrueger.de> wrote:
> No answer at all to my question?
>
> Is the wrong number of estimatedResultCount by AjaxApi perhaps a
> sinister Google strategy to make programs like mine work less
> effective?
>
> As a workaround,
> I could quite as well send a normal website-google-request,
> and then parse the results page for that number,
> only that is more work than necessary, I find.
>
> Please say something.
>
> Please have a look at my measurements:http://www.andreaskrueger.de/google/
>
> Thanks a lot!
> Andreas
>
> On Jul 11, 2:12 am, AndreasKruegerDe <gro...@andreaskrueger.de> wrote:
>
> > There seem to be inconsistencies between the WebSite results,
> > and the Ajax results, concerning the __EstimatedResultCount__
>
> > Depending on the question,

> > the (manual)www.google.comsearchresults' 'estimatedResultCount'

> > are up to about 10 times higher than the Ajax
> > 'estimatedResultCount' !!!
>
> > And it is not even always higher, sometimes they are smaller!!!
>
> > Please have a look at my measurements:http://www.andreaskrueger.de/google/Google_EstimatedResultCount_Websi...
>

[Ben Lisbakken]

Hello --

Unfortunately, the estimated result count is a rough estimate and is not reliable enough to be used for research/exact calculations.  It is subject to frequent change as our index changes.  Also, there are no plans to change its behavior.

Thanks,

Ben Lisbakken