Insights from filter toggling

more experimentation at Google?

[Slud]

It used to seem that appending "&filter=0" had two effects:
1) uncap the # of results that could show up from a single domain
2) allow near-duplicate pages to be displayed

I just tried it on a query that was giving particularly less relevant results, and got different results showing up (not just new results stuck in the listings). In fact the results looked much more like last month's results (i.e. less random).

Is it possible that the similarity filter is now one of degree instead of a binary toggle?

I don't think I can share they keyphrase here, so sticky me if you're interested.

[vitaplease]

Slud,

could you elaborate on this "&filter=0" or sticky me an example?

[ciml]

Anyone else find differences when appending &filter=0 to the URL of the results page (other than more than two results per domain)?

[Slud]

Unfortunately, I can't find any more examples of this after the Oct. 31 update (thank god).

My mind still boggles at how random the results were last month.