3 ms·

I'd like to see some experiments with Search engines that rank on things that are hard or impossible to game. Impossible in the sense that the search-engine con

by berkes 3y ago

I'd like to see some experiments with Search engines that rank on things that are hard or impossible to game. Impossible in the sense that the search-engine controls this variable entirely. Like "age of the article" through "time we saw it first & how much it changed since". Or like "churn", through "how often we saw new content come and go". Anything that users or owners control, will be gamed. So maybe there are variables that only the search-engine itself controls that still provide adequate or good ranking?

halflings 3y ago
But that is exactly what search engines do. Yet people find a way to game these "impossible to game" metrics. For example: "age of the article": search engines value recent content => suddenly you start seeing articles published just weeks/months ago reviewing some rather old piece of hardware/content. Either a full repost under a different URL with a different title etc., or just an incremental (probably automated) update of an older page.
- berkes 3y ago
  I probably wasn't clear. I truly meant "impossible to game". If I crawl your site today, encounter article A, and crawl it again in a year and again encounter article A, I am sure that article A is a year old (Well, techcnically, what I am certain of, is that I encountered the exact same article two times a year apart). You cannot game that. Now if another year later I encounter article A' where some content on the page has changed, be it sidebars, a new design, or a few paragraphs, I might say, "some version of this article is two years old" and decide to rank that different from "this exact version is two years old". Age is just one variable that I can think off. I'm quite certain there are more variables. But each and every one such variable must be completely controlled and verified by me- the search engine. Nothing external can be such a variable: not "amount of articles that link to X" nor "amount of people that click through". Maybe such variables don't exist. Or maybe they do, but make for terrible ranking, IDK. But I do know that using variables that search-engines don't control make for a terrible search experience over time. As can be witnessed and as is written down in the linked article.
  - CJefferson 3y ago
    Favoring old articles might work, but it would wouldn’t work for any articles written more than 12 hours after SEO types find out it’s an important ranking.
    - berkes 3y ago
      Why not? It's not as if SEO types have time machines?
  - jabradoodle 3y ago
    > If I crawl your site today, encounter article A, and crawl it again in a year and again encounter article A, I am sure that article A is a year old (Well, techcnically, what I am certain of, is that I encountered the exact same article two times a year apart). You cannot game that. And what if I change a tiny amount of the article, or post it on a second domain I own? Do you think you can match up all the near duplicates at the scale of the internet?
    - NL807 3y ago
      Perhaps that's ok, because it's about persistence of information, not where it lives. I suppose you could view this as longer some information lives somewhere (doesn't matter where) the higher ranking it gets. If that information gets relocated, not a problem, it was seen somewhere else before, therefore it could be still regarded as high ranking.
      - jabradoodle 3y ago
        That is my point, unless the content is identical, any tiny change and I don't see how your going to know it is the same page as another when you have billions/trillions of pages to compare against. I also don't think this makes sense on a fundamental level, if I write some excellent content it shouldn't be harder to find because I wrote it recently. We would be making it harder for new entrants and encouraging people to rush out low quality content/ even report on things before they happen and delete the incorrect guesses later (not that the current system isn't full of perverse incentives.)
beej71 3y ago
I'd like to see it possible to game search engines by having the fewest ads on the page. :)

by berkes 3y ago

halflings 3y ago

But that is exactly what search engines do. Yet people find a way to game these "impossible to game" metrics. For example: "age of the article": search engines value recent content => suddenly you start seeing articles published just weeks/months ago reviewing some rather old piece of hardware/content. Either a full repost under a different URL with a different title etc., or just an incremental (probably automated) update of an older page.

berkes 3y ago
I probably wasn't clear. I truly meant "impossible to game". If I crawl your site today, encounter article A, and crawl it again in a year and again encounter article A, I am sure that article A is a year old (Well, techcnically, what I am certain of, is that I encountered the exact same article two times a year apart). You cannot game that. Now if another year later I encounter article A' where some content on the page has changed, be it sidebars, a new design, or a few paragraphs, I might say, "some version of this article is two years old" and decide to rank that different from "this exact version is two years old". Age is just one variable that I can think off. I'm quite certain there are more variables. But each and every one such variable must be completely controlled and verified by me- the search engine. Nothing external can be such a variable: not "amount of articles that link to X" nor "amount of people that click through". Maybe such variables don't exist. Or maybe they do, but make for terrible ranking, IDK. But I do know that using variables that search-engines don't control make for a terrible search experience over time. As can be witnessed and as is written down in the linked article.
- CJefferson 3y ago
  Favoring old articles might work, but it would wouldn’t work for any articles written more than 12 hours after SEO types find out it’s an important ranking.
  - berkes 3y ago
    Why not? It's not as if SEO types have time machines?
- jabradoodle 3y ago
  > If I crawl your site today, encounter article A, and crawl it again in a year and again encounter article A, I am sure that article A is a year old (Well, techcnically, what I am certain of, is that I encountered the exact same article two times a year apart). You cannot game that. And what if I change a tiny amount of the article, or post it on a second domain I own? Do you think you can match up all the near duplicates at the scale of the internet?
  - NL807 3y ago
    Perhaps that's ok, because it's about persistence of information, not where it lives. I suppose you could view this as longer some information lives somewhere (doesn't matter where) the higher ranking it gets. If that information gets relocated, not a problem, it was seen somewhere else before, therefore it could be still regarded as high ranking.
    - jabradoodle 3y ago
      That is my point, unless the content is identical, any tiny change and I don't see how your going to know it is the same page as another when you have billions/trillions of pages to compare against. I also don't think this makes sense on a fundamental level, if I write some excellent content it shouldn't be harder to find because I wrote it recently. We would be making it harder for new entrants and encouraging people to rush out low quality content/ even report on things before they happen and delete the incorrect guesses later (not that the current system isn't full of perverse incentives.)

beej71 3y ago

I'd like to see it possible to game search engines by having the fewest ads on the page. :)