News for hackers. Stuff that matters.
Re:Google Web spam - Gabriel Weinberg's Blog by moultano on Tuesday July 27, 2010 @06:24PM (#1552939)
Being in the index isn't very meaningful. Very little of our spam-fighting/ranking prevents sites from showing up for "site:" queries, because in general we think that if someone is intending on going to a domain directly, the only reasonable thing to do is to show that domain.

Unfortunately I don't have a better way of assessing it to offer. Internally, we often look at impression-weighted precision as a metric, but I don't think there's an easy way we could expose that to you.

A more reasonable thing to do would be to take a sample of DDG's query logs, scrape the results from Google, then see what percentage of Google's results come from your spam domains, but that requires sending a lot more queries to get any useful data.

Re:Google Web spam - Gabriel Weinberg's Blog by epi0Bauqu on Tuesday July 27, 2010 @06:28PM (#1552950)
I think this method would violate (at least the spirit of) my privacy policy: http://duckduckgo.com/privacy.html

That said, if anyone has a meaningful sample query set, I'm certainly interested in running it against my spam index. I see a lot of hits on it via other search APIs.

Re:Google Web spam - Gabriel Weinberg's Blog by gojomo on Tuesday July 27, 2010 @06:31PM (#1552957)
You could let people opt their queries into such studies.
Re:Google Web spam - Gabriel Weinberg's Blog by epi0Bauqu on Tuesday July 27, 2010 @06:39PM (#1552982)
I don't have accounts right now so it's a bit tricky. If there was some way people could export their Google search history, then I could use that. Feel free to email it in anonymously.
Re:Google Web spam - Gabriel Weinberg's Blog by moultano on Tuesday July 27, 2010 @07:17PM (#1553050)
I found this link which looks like it dumps search history as an rss feed. That might be a convenient way to send it over.

https://www.google.com/history/?lookup?q=&output=rss&...

Re:Google Web spam - Gabriel Weinberg's Blog by pierrefar on Tuesday July 27, 2010 @07:03PM (#1553025)
Can you please explain "impression-weighted precision" and how it's used? Sounds very interesting.
Re:Google Web spam - Gabriel Weinberg's Blog by moultano on Tuesday July 27, 2010 @07:10PM (#1553039)
In general we want to measuring what matters to users, so removing a lot of spam that nobody sees doesn't really make anything better.

By the same token, if you are trying to launch a new spam classifier that has some false positives, if one of the false positives is yahoo or facebook, it doesn't really matter how good it is, it will never be worth the collateral damage.

As a result, rather than measuring precision/recall as a percentage of domains or as a percentage of urls we usually try to measure it as a percentage of results that appeared on a search result page or results that users click on, mined from our logs.

This is one of the bajillion reasons why it's absurd to expect Google to throw away all its logs data. The logs are essential to coming to any meaningful conclusions about the current quality of our search, let alone finding ways to improve it.

Re:Google Web spam - Gabriel Weinberg's Blog by pierrefar on Tuesday July 27, 2010 @07:17PM (#1553049)
Thanks.

Am I right in understanding that if a SERP result for a given keyword doesn't get clicked by users enough, it will be removed?

By the same token: does the result's bounce rate matter? I imagine spammy sites have a very high bounce rate.

Re:Google Web spam - Gabriel Weinberg's Blog by ddemchuk on Wednesday July 28, 2010 @12:25PM (#1554742)
"As a result, rather than measuring precision/recall as a percentage of domains or as a percentage of urls we usually try to measure it as a percentage of results that appeared on a search result page or results that users click on, mined from our logs."

So this basically means that you are able to discern content spam on authoritative domains (facebook, wordpress.com, etc) based on ctr, bounce, impressions compared to surrounding serp results rather than comparing that data against the parent domain as a whole?

Re:Google Web spam - Gabriel Weinberg's Blog by moultano on Wednesday July 28, 2010 @12:26PM (#1554749)
No, not at all. This is just talking about the granularity at which we compute metrics like "spam rate."