Forum Moderators: open
Why do some results appear for widget -widget?
For rarely entered search terms such as "bigwidgets -bigwidgets" or "foo -foo" (foo = 200,000 pages) nothing appears.
However for more popular terms doing <popularterm> -<popularterm> often a few hundered terms appear. Any suggestions as to why?
So I thought perhaps...
1. Google was taking note of the most popular search terms (as oppossed to words/phrases no one really searches for) - and using some unique algorithm only on these. Nope.
2. I then noticed that several of these results contained keywords in the navigation bar - perhaps google ignores the html that stays the same from one page to the next. Then I noticed that in some of the results the keywords only appear in spam lists - e.g. "cheap widgets widgets for sale big widgets". So perhaps we're seeing the results of an algorithm that Google uses to decide which parts of a page can be ignored. Nope.
3. I then saw results containing the keywords inside normal text. On closer inspection many of these had <font>keyword</font> or similiar things surrounding the keyword. Nope.
...and tried a few more things that I can't remember!
[google.com...]
hence the handful of pages returned for "the bar" -"the bar"
wonder why the #99 (Google Toolbar-toolbar.google.com) shows when there is not a single 'bar' on that page?
I searched for yahoo -yahoo [google.co.uk]
I noticed that at the top of the page google has the prompt "do you want to return results in english only?". Doing so returns no results (even though there are definitely english language pages in the results).
The same is also true of "the bar" -"the bar" (although there is no prompt, there are no results in an English only search).
Perhaps the results for widget - widget are due to Google's language features in some way? Maybe they are translating the term in the serps or something?
Maybe the "widget -widget" technique only removes pages from the SERPs that have been guaranteed/confirmed to be in English, even though lots of the returned pages are in English.
A reason for this could be the word "widget" has another meaning in some language other than English - and Google does not want to exclude pages about the other meaning of the word - maybe cause the '-'-algorithm only works for typical English words.
Anyone still following me? Using the "widget -widget" query on typical Dutch words which aren't used in any other language (e.g. telefoon, haardroger) confirms this, as there are no results at all.