Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Dear Google: please let me ban sites from results

news.ycombinator.com

Post reply on HN
Print view View original

121–130 of 217 posts

Re: Dear Google: please let me ban sites from results

#121

Post by nervechannel »

AndrewO wrote: post #104

I see a lot of people asking what happens when a group of people downvote a site just to ruin its ranking. Sure that's a problem, but there's an easy solution on Google's end: your blacklist only affects you. Yes, that means all of us have to hide efreedom ourselves. Doesn't seem like a problem to me... Plus, we are talking about a company whose core business demands that it can identify groups of bad-faith voters. G…

This is exactly what I meant. Lots of people have added the idea about crowd-sourced re-ranking based on blacklisting, then said "it'll never work..."

Re: Dear Google: please let me ban sites from results

#122

Post by djhworld »

ergo98 wrote: post #44
djhworld wrote: post #20

I think the worst culprits are the ones that skim StackOverflow questions and rehash them into their own supposed original "question and answer" site

What do you think most StackOverflow answers are? It's a karma-paid labor pool where you can post questions and a lot of under-employed people will rush out and do the necessary Google searches, collating and slightly rewriting the results to yield the most votes. Everyone is ripping off someone's content. And just to be accurate here, SO content is creative commons (created by the community). Are those just cheap wo…

It's not the answers I'm bothered about, it's just how they spam my search feed. I whack in some queries to google and the first 10 results will just be the same SO question but spread over multiple different sites.

Re: Dear Google: please let me ban sites from results

#123

Post by docgnome »

eli wrote: post #53
al_james wrote: post #7

Yes that would be good. They could then look at the number of people blocking certain domains and de-weight them in the global results. Traditionally google seem against human powered editing (as this would be), but I think as the black hat SEOs run rings around them, its needed badly .

An extremely easy way for a bunch of people to get together and destroy someone's ranking? That doesn't sound like such a good idea.

4chan would go bananas.

Re: Dear Google: please let me ban sites from results

#124

Post by joshrule »

It seems that it might be more helpful to whitelist sites. The web grows too quickly, and the mass of spam sites overwhelmingly so. If I had some way to blacklist sites, I'd end up spending a lot of time doing so. In fact, it could quickly take up most of my search time.

If, though, we could whitelist sites, it seems that results would get cleaner faster. I don't care about how many bad sites are out there, as long as helpful sites make it to the top. Plus, I typically use just a few sites to access reliable information anyway (the number's about 7, right?), so if I can whitelist results from those sites, I'll probably find my desired content more quickly.

What about the case when there are 30 spam sites listed before 1 good site? That hasn't happened too often for me. Instead, the results I'm looking for are usually just 4 or 5 spots down the front page, and very occasionally on the second page.

White listing seems like it would still be faster and easier for now.

Re: Dear Google: please let me ban sites from results

#125

Post by bad_user »

troutwine wrote: post #115
seabee wrote: post #89

Earlier quoted context omitted.

Do you think this could be improved by alternate wording? Instead of reporting 'spam', ask the user 'was this useful?' or 'did you want this?'. After all, the real goal is giving people a better, more relevant experience, detecting and removing spam is just one facet of that. Whether it's email or search.

I don't think changing the wording will solve the problem entirely. On the rare occasion that my wife receives spam via gmail she'll simply archive it. The action to hide an email you never did or no longer do care about is so common and so easily accessed by a well-learned key cut that it doesn't occur to her to do otherwise until after the fact; she is well aware that marking spam improve spam filtering, of course.

I just told my wife: think of that button as being "Report Abuse".

She's been applying it ever since quite efficiently.

Re: Dear Google: please let me ban sites from results

#126

Post by peterhoffmann »

AndrewO wrote: post #104

I see a lot of people asking what happens when a group of people downvote a site just to ruin its ranking. Sure that's a problem, but there's an easy solution on Google's end: your blacklist only affects you. Yes, that means all of us have to hide efreedom ourselves. Doesn't seem like a problem to me... Plus, we are talking about a company whose core business demands that it can identify groups of bad-faith voters. G…

The blacklist does not only have to affect me, just throw in the blacklists of my social graph too. These are the people I trust.

Re: Dear Google: please let me ban sites from results

#127

Post by jawn »

moe wrote: post #109
tomjen3 wrote: post #79

Earlier quoted context omitted.

Most botnets would easily have more than that number of computers. So yeah it can be done.

It would also be a significant undertaking if the down-votes are tied to google accounts. Not only do you have to bypass the captcha, but also the heuristics that google could (does?) employ to detect fake accounts. If the votes were weighted by account activity then the spammers would not only have to sign up and vote; they would also have to keep all these accounts busy with fake queries or other means.

Gmail accounts and capcha solving services are cheap compared to the profits you could make off of getting your site ranked #1 for a profitable search keyword.

Re: Dear Google: please let me ban sites from results

#128

Post by jasonkester »

travisp wrote: post #96
bradly wrote: post #71

Earlier quoted context omitted.

And ExpertsExchage and About.com

ExpertsExchange is at least original content (as far as I know)

ExpertsExchange used to be nearly 100% scraped usenet. Then eventually a few poor souls stumbled in thinking it was an actual site and started posting questions there directly. Eventually people even started answering them.

Had they not had such a terrible site design and revenue scheme, they might have actually made the transition from scraping bottom feeder to respectable site.

Re: Dear Google: please let me ban sites from results

#129

Post by AndrewO »

nervechannel wrote: post #121
AndrewO wrote: post #104

I see a lot of people asking what happens when a group of people downvote a site just to ruin its ranking. Sure that's a problem, but there's an easy solution on Google's end: your blacklist only affects you. Yes, that means all of us have to hide efreedom ourselves. Doesn't seem like a problem to me... Plus, we are talking about a company whose core business demands that it can identify groups of bad-faith voters. G…

This is exactly what I meant. Lots of people have added the idea about crowd-sourced re-ranking based on blacklisting, then said "it'll never work..."

It seems to be a particular affliction among our group (by that I mean anyone who spends a lot of time here) that once we learn to apply one revolutionary/disruptive idea, we can't stop even when we probably should. I'd say it's because we're used to trying to think of how to scale-up every idea, but I've been accused of armchair-psychology before... :)

Re: Dear Google: please let me ban sites from results

#130

Post by bad_user »

bromley wrote: post #119
moe wrote: post #99

Earlier quoted context omitted.

For a keyword that only gets 100 searches a day, it wouldn't take many down-votes to affect the rankings of the relevant sites. Why do you assume a flawed implementation? Naturally there would be thresholds. There is no reason to devalue a site that's only displayed in 100 search-results/day at all. The sites we want to hit are orders of magnitude worse at polluting the results. We're talking about the Mahalo's and e…

Google seems to like building systems that automatically scale across all keywords and niches. For example, for a low competition keyword, 1 link can make a big difference to a site's rankings. For a high-competition keyword it takes many many more links to change anything. It's the same algorithm affecting the rankings in both cases, just in the latter case there's a lot more data being used as input. That's the way…

    Google seems to like building systems that automatically
    scale across all keywords and niches
They haven't been doing a good job lately.

    For a high-competition keyword it takes many many more 
    links to change anything.
Not that many, and do you think that's a problem?

Search "viagra": http://www.google.com/search?q=viagra

This is the third result (after wikipedia + viagra.com which have a hardcoded boost): http://www.genericviagrarx.net/

Here's a sample of links to it: http://www.google.com/search?q=link:genericviagrarx.net

I.e. there are thousands of forums / blogs where you can post your link to (even in a proper context, the more popular the subject the better). Hire a couple of offshore workers for $5 / hour, and in a week you'll have thousands of links pointing to you.

Post reply on HN