How does Google count and estimate the number of a search results?

Question

How does Google count and estimate the number of a search results? For example when I search "stackoverflow," it counts 2,910,000 results.

Answer 1

Probably the most definitive article on this:

http://searchengineland.com/why-google-cant-count-results-properly-53559

as it's interesting to note that taking stuff out can actually sometimes give more results(!!).

From a Google developer (Matt Cutts, head of the web spam team):

"We try to be very clear that our results estimates are just that--estimates. In theory we could spend cycles on that aspect of our system, but in practice we have a lot of other things to work on, and more accurate results estimates is lower on the list than lots of other things"

Answer 2

I believe this question is technical than conceptual.

And that is MEMCACHE.

To my knowledge, memcache, does any operation in a standard amount of time. immaterial of the amount of records. Adistributed RAM architecture. For a similar realtime implementation, read Sharding Counters article. They keep clusters containing of about 10-1000 2-4 gb RAMs, They do temporory math in them, When they persist the final value, they only sum up the sums in each of the RAM.

My use of that implementation tells it will be faster (and accurate too). It is open source now, visit memcached.

Coming to the point that says the company 'estimates' values... My view is big companies (including google) divert people from truth. A lot many things are dumped on the people so that people will only use what they are offered, and don't become competitors to companies in the near future at the least.. People means companies, investors, talent potential techies, everyone.

If they estimate, there are so many things they do which are not possible on 'estimation'. Like wallet api, adsense invoicing, etc.