Does tag spamming actually change how the indexer ranks your project? Or is it just a myth?

From what I understand, putting any tag will allow the project to be searchable under that tag, and also puts it under certain relevant pages on the Explore page.

As far as I'm aware, there is no concrete evidence that spamming the same tag multiple times is any better than just putting the tag once. This is a popular theory, but from what I see it mainly relies on the “correlation implies causation” fallacy.

Interesting. It would probably not be allowed to test this though, correct?

SpyCoderX wrote:

Interesting. It would probably not be allowed to test this though, correct?

I'm not sure how you'd test it. Suppose that you make a project, tag spam on it, and then it gets popular. How would you be able to tell that it was the tag spamming that made it popular? Maybe it was just a really good project.

Za-Chary wrote:

SpyCoderX wrote:

Interesting. It would probably not be allowed to test this though, correct?

I'm not sure how you'd test it. Suppose that you make a project, tag spam on it, and then it gets popular. How would you be able to tell that it was the tag spamming that made it popular? Maybe it was just a really good project.

You could test it by making the actual project just a test and nothing of good quality (or better yet, assume it is real)

ajskateboarder wrote:

Za-Chary wrote:

SpyCoderX wrote:

Interesting. It would probably not be allowed to test this though, correct?

I'm not sure how you'd test it. Suppose that you make a project, tag spam on it, and then it gets popular. How would you be able to tell that it was the tag spamming that made it popular? Maybe it was just a really good project.

You could test it by making the actual project just a test and nothing of good quality

Yea, that’s probably what I’d do. And then use a tag like “#nothing” and make a second project with only 1 #nothing tag.

The thing I’m mainly worried about is that I’d get banned for it.

Za-Chary wrote:

From what I understand, putting any tag will allow the project to be searchable under that tag, and also puts it under certain relevant pages on the Explore page.

As far as I'm aware, there is no concrete evidence that spamming the same tag multiple times is any better than just putting the tag once. This is a popular theory, but from what I see it mainly relies on the “correlation implies causation” fallacy.

I recommend reading this and the document from the ST in this post. Also, I once asked @Codubee if he was aware of the bug and, from what I remember, he replied that he was (it was years ago)

Za-Chary wrote:

SpyCoderX wrote:

Interesting. It would probably not be allowed to test this though, correct?

I'm not sure how you'd test it. Suppose that you make a project, tag spam on it, and then it gets popular. How would you be able to tell that it was the tag spamming that made it popular? Maybe it was just a really good project.

Compare the growth speed to another “control” project, both projects would be the same, but one would use 1 of each tag and the other would spam tags. Repeating this multiple times to make sure that it's the tag Spamming causing the faster growth.

ajskateboarder wrote:

Za-Chary wrote:

SpyCoderX wrote:

Interesting. It would probably not be allowed to test this though, correct?

I'm not sure how you'd test it. Suppose that you make a project, tag spam on it, and then it gets popular. How would you be able to tell that it was the tag spamming that made it popular? Maybe it was just a really good project.

You could test it by making the actual project just a test and nothing of good quality (or better yet, assume it is real)

It would be better to use a full project, maybe a generic platformer simple game, as it would be more likely to get on trending, because a generic platformer Simple game would probably be liked more than a blank screen, which would make it easier to spot differences in growth speed.

SpyCoderX wrote:

ajskateboarder wrote:

-snip-

The thing I’m mainly worried about is that I’d get banned for it.

It's not possible to be banned for tag spamming [citation needed], it's not allowed, yes, but it's not bannable. The worst case is the project being taken down and a temporary block. Even that is unlikely, it's possible that the project is taken down, but the block is unlikely.

CST1229 wrote:

It is most likely real.

If any of this was confirmed by the ST, it would mean that tag Spamming is a functional exploit. To add to this ;

Paddle2See wrote:

Yeah, that's a bit of an abuse of the system that will eventually make it useless if it continues. It's also against the Community Guidelines - the part about “being honest”. Furthermore, it's also against the Terms of Use - the part that talks about interfering with the service (3.5). So feel free to link anybody you see doing that to this topic so they become better educated. You may also use the Report button and explain that they are abusing tags.

(source)

Based on this, tag spamming is most likely a functional exploit.

Your not supposed to do it, real or fake.

alexandretherrien wrote:

I recommend reading this and the document from the ST in this post.

Thanks; I was not previously aware of the document. This now seems like solid proof (at least at a glance).

Za-Chary wrote:

SpyCoderX wrote:

Interesting. It would probably not be allowed to test this though, correct?

I'm not sure how you'd test it. Suppose that you make a project, tag spam on it, and then it gets popular. How would you be able to tell that it was the tag spamming that made it popular? Maybe it was just a really good project.

There is a way to test the bug that works 100% of the time. If you spam tags or use any of the other tag related bugs, your project will always go to the top of the recent page for the explore tab that goes with tag you used. So, if you spam in the note and credit the tag #animations, your project will go 100% of the time to the top of this page. This is because the recent page does not take project popularity into account when showing project. So project shown are just projects where the right tags are being used. So, when a project abuse a certain tag, it always go to the corresponding recent page on explore.

I've done this test in the past, but to prove it, yesterday I made 2 projects that I shared at the same time with 0 views each. One has like 100 times the tag ‘'#art’' and the second one has one time the tag ‘'#art with other text so that the note and credit section is not just one tag. As expected, only a few hours later the project with spammed tag got the first spot on ’'recent'' for the Art category while the other test project is no where to be seen. Btw if someone does the test now, both projects need to have a low id because high IDs are still not indexing. 30 million projects are missing

alexandretherrien wrote:

-snip-

Well then, I guess someone beat me to testing it heh. That is definitely abuse of the system lol (using tag spamming).

I hope the scratch team will patch it at some point, although the indexer seems like a bigger issue.
Theoretically, the tag spamming could be fixed if the indexer just removed all words starting with # when it does the scans for keywords (art, animation, etc.).

Thanks for testing it for me though lol

SpyCoderX wrote:

(#13)
Theoretically, the tag spamming could be fixed if the indexer just removed all words starting with # when it does the scans for keywords (art, animation, etc.).

Probably not, since that would remove all actual tags, and wouldn't prevent just typing the word itself without a hashtag to make it show in the explore tabs.

CST1229 wrote:

SpyCoderX wrote:

(#13)
Theoretically, the tag spamming could be fixed if the indexer just removed all words starting with # when it does the scans for keywords (art, animation, etc.).

Probably not, since that would remove all actual tags, and wouldn't prevent just typing the word itself without a hashtag to make it show in the explore tabs.

Hmm. Well, perhaps just remove that functionality and have a separate section where you can add (a limited number of) tags? (I'll probably make a suggestion about this).