Google result counts

There has been plenty of speculation on the web over the years about the result counts listed from Google searches. I have my own bit of interesting Google search result trivia.

Before TurboGears’ release, searching for the word turbogears would yield a few dozen automotive-related hits, which makes sense because it wasn’t really a word or a brand name before now.

This morning, a search for turbogears yields 210,000 results. That number is so large that I find it hard to believe that even counting automatically-generated archives of things will yield anywhere near that number of pages. So, I decided to jump through the results. Google’s results took me to “981-984 of about 210,000”. Even on the *last page* of results, it was still saying 210,000. Odd, huh?

While I know that Google has always claimed that the result counts are estimates, I wonder how they compute the estimate… Even Yahoo seems inflated at 15,600.

Maybe Google’s just out to make vanity searchers happy 🙂

3 thoughts on “Google result counts”

  1. They probably have a separate index bucket for keywords yielding the amount of results that “turbogears” seems to generate (also considering how fast the number of matches this keyword gets has been growing lately) and so it comes up with such an estimate to all keywords that fall on this same bucket.

    On a side note, I wrote several months ago a quick and dirty python script to query a predefined set of keywords and store the amount of results google guesses, as a way to track the growth of hits on google for these specific keywords. It’s available at
    http://rnc000.blogspot.com/2005/09/google-hits.html
    (you may need gnuplot to draw the graphs from the generated data and a daily cron job)

Comments are closed.