[IvyMike]

> Most numbers of this magnitude garner a few thousand hits on Google, but 25898913 gets 29,500,000.

If you actually click through the results, on pages 1-34 you get "Page 34 of about 24,900,000 results (0.34 seconds) ", but on page 35 you discover that there are only actually 341 results.

[ajani]

I've found this to be consistent behaviour with Google. The number of search results tend to be totally unreliable. I don't know if they use the same search algorithm for gmail, but the same problem occurs there as well. Initially it might say 1-50 of 100 results, go to the second page and suddenly it's 50-100 of 2000. And on the next page 100-150 of "many".

Google search-hit count has never been fully accurate - I say this empirically from use over the years, and having tried to use the number of hits in some random experiment I can no longer remember.

[vtlynch]

Google acknowledges this: https://support.google.com/gsa/answer/2672285?hl=en

I remember seeing a more in depth article about this but cant find it at the moment.

[Chinjut]

So apparently it is possible to get an accurate result count (for up to 1 million results) using "rc = 1"? Does anyone have an examples illustrating this behavior (I can't seem to contrive one where it makes a difference)?

[ajani]

I suspect rc=1 applies to enterprise search.

[Chinjut]

Ah, that makes sense. Thanks!

[arikrak]

It's a bit ridiculous how Google shows those numbers on top and people constantly quote there are X results for Y when the real results are always far off from that. There are 382,000 results for {number of "google results"}!

[FreeFull]

Sounds like an odd bug in how Google works.