Thanks for providing these insights. Russia has remarkable computer scientists and engineers. Could be a glimpse into the future of search.

Google has experimented with this and reportedly they feel the link signal still works best. But it seems like they're bolting on a lot of anti-spam and user intent filters to keep it working. In my mind's eye the algorithm resembles a chugging steampunk machine with disparate and unplanned parts bolted on over time to extend it's function as needs arised. I don't think its far fetched that an alternative to links could arrive in five to ten years.

i think a search engine relying primarily on links is always second best and must finally lose compared to search engines who successfully introduce other ranking factors.

natural links you have to earn over time. so first of all there is always a time factor which hinders actuality of results. on google, old websites have always an unjustifyable advantage over new websites. if you have a new website with superior quality and usability, there's almost no way to show that to google.

best way to drive natural links for me has been heavy exposure on the search engines. but if you don't have a well-established site, there's no way to earn these links in an adequate time frame so that's going to be a non-starter.

without links, no ranking - without ranking, no links. that's the vicious circle for newcomer websites on google which drives me nuts.

I think Yandex has the right idea: Pay attention to links when they're likely to be true citations (i.e., in informational searches) and ignore links that are likely to be SEO-driven.

i think a search engine relying primarily on links is always second best

Ultimately it's an admission of defeat, isn't it? "Frankly we can't tell what makes a good site, so we'll just count how many other sites link to it and rely on their judgement."

Ultimately it's an admission of defeat, isn't it?

In 1998 it was an innovative idea and an astute use of available data.

And Google used it to help wipe the floor with its competitors.

we'll just count how many other sites link to it and rely on their judgement

It's worse than that - "Google’s algorithms rely on more than 200 unique signals".

So there are more than 199 other admissions of defeat yet to be exposed.

...

we'll just count how many other sites link to it and rely on their judgement

It's never been that simple. (Not even remotely.)

Yes the "how many sites" was never a ranking determination in Pagerank, but link popularity was used as a good rough guide for when you didn't know the PR of a page.

Thanks for sharing the original post, sounds like ingenuity is hard at work.

Yandex is good at analysing landing pages, it understands what type of a landing page the visitor expects to see within a certain niche.

The problem with this is that (not unlike similar issues with Google) only sites that fit a certain pattern can rank well. So innovative sites, that real humans love if they get to see them, sink, while uniformity reigns. :(

Sites publishing Yandex Direkt ads (alternative to Google Adsense) seem to have a priority

That would be a very serious matter if it were shown to be true.

do you think Bing or Google might attempt something similar?

Only if they have a death wish.

...

If this experiment is successful, do you think Bing or Google might attempt something similar?

Probably (I'm sure they're watching the Yandex experiment closely), but an equally interesting question is:

"If Google or Yahoo did devalue links as a ranking signal in commercial searches, how long would it take them to reveal that?"

From a search-engine Webspam team's point of view, keeping SEOs and site owners in the dark might have a benefit: If the owner of Cheapo-payday-loans-and-herbal-drugs dot com is madly buying links that won't accomplish anything, he'll have less time and money for more productive forms of spamming.