Discuss Scratch

PaddlerGames
Scratcher
1000+ posts

Is trending weirdly subjective?

RokCoder wrote:

I've recently released two projects that have done reasonably well yet neither of them appear in the trending pages at all (and I've looked back through many, many pages out of curiosity). This is fine but what makes me wonder about the trending algorithm is when searching for the project by name and then sorting it by trending, my own projects are featured well below the remixes. Slightly disappointing to say the least

This is the link to the sorted search - https://scratch.mit.edu/search/projects?mode=trending&q=uno%20deluxe

The project is getting maybe 150 likes per day so I would have expected the algorithm to at least be putting it above the remixes

Does anyone have any insight into this? Maybe a member of the ST?

—–

My browser / operating system: MacOS Macintosh X 11.0.0, Chrome 86.0.4240.198, No Flash version detected
When a project gets 100+ loves in its first 24 hours, it is blacklisted from appearing on trending. It seems that way at least. I've been noticing this has been happened since June with some popular scratchers projects, such as griffpatch, sharkyshar, huntedskelly, etc. so I made a forum topic about it: https://scratch.mit.edu/discuss/topic/454296/ It seems like the only way to bypass this rule is to reshare your project, which is why when you reshared your Uno game, it got on trending.
CloudyWhale
Scratcher
64 posts

Is trending weirdly subjective?

Za-Chary wrote:

Yeah, I have no idea how Trending works. You've seen my recent game, “Topological Space,” correct? That was Top Loved and Top Remixed (it kinda jumped back and forth between the two sections) — but throughout that time, it was nowhere to be found on Trending. I could only suspect that the Trending algorithm was changed recently (I don't know that for sure). If so, did it change for better or for worse? I suppose that would be up for debate.

At this point I have the mindset of “As wonderful as Trending can be sometimes, we should just not worry about it.” There has been quite a lot of drama and disrespect that was caused by certain projects getting on Trending and certain projects not getting on Trending… at least for my sake, it's best to stay away from it.

Try asking your fellow scratch team members if you are curious

And the algorithm needs to be fixed people make cloud scrolling platformers with tags in the title because that is what the algorithm always picks for some reason. I'm tired of seeing these. Please Stop fighting for fame.

Edit: I really never go to the trending page because i am sick of seeing those projects that basically USE THE ALGORITHM FOR FAME.
If you go to https://scratch.mit.edu/projects/417586431/ it has lots of tags i believe it was made to show how the scratch algorithm is messed up it got 1st on trending super fast

Last edited by CloudyWhale (Dec. 4, 2020 19:55:09)

83ben38
Scratcher
13 posts

Is trending weirdly subjective?

PaddlerGames wrote:

RokCoder wrote:

I've recently released two projects that have done reasonably well yet neither of them appear in the trending pages at all (and I've looked back through many, many pages out of curiosity). This is fine but what makes me wonder about the trending algorithm is when searching for the project by name and then sorting it by trending, my own projects are featured well below the remixes. Slightly disappointing to say the least

This is the link to the sorted search - https://scratch.mit.edu/search/projects?mode=trending&q=uno%20deluxe

The project is getting maybe 150 likes per day so I would have expected the algorithm to at least be putting it above the remixes

Does anyone have any insight into this? Maybe a member of the ST?

—–

My browser / operating system: MacOS Macintosh X 11.0.0, Chrome 86.0.4240.198, No Flash version detected
When a project gets 100+ loves in its first 24 hours, it is blacklisted from appearing on trending. It seems that way at least. I've been noticing this has been happened since June with some popular scratchers projects, such as griffpatch, sharkyshar, huntedskelly, etc. so I made a forum topic about it: https://scratch.mit.edu/discuss/topic/454296/ It seems like the only way to bypass this rule is to reshare your project, which is why when you reshared your Uno game, it got on trending.

Makes sense, but then reshapes get 100+ loves in the first 24 hours technically…
12345catc0de
Scratcher
100+ posts

Is trending weirdly subjective?

I think trending page shows projects that has a lot of views and I mean A LOT. Take my Starter project remix for example, it has about 2000 views so that is why it got #90 on Trending.
pranavhaha
Scratcher
91 posts

Is trending weirdly subjective?

When it is going to fix? I will ask @paddle2see

Last edited by pranavhaha (Dec. 5, 2020 10:48:49)

TheCloudDev
Scratcher
100+ posts

Is trending weirdly subjective?

-moonchild wrote:

MeIzAwezomeDede wrote:

-moonchild wrote:

RokCoder wrote:

snip

I believe one of the contributing factors to this would be that you might have too many followers? I saw a few people that have a large amount of followers say they won't be getting on trending anymore since they have that large amount of followers. I'm not sure if that was what you were speaking about.
I don't know if that's why, there's plenty of famous people on trending right now

That's odd then, maybe there's a specific amount of times people can be on trending?
I don't think so……. we have al seen project b the same user come on trending QUITE a few times….
Chiroyce
Scratcher
1000+ posts

Is trending weirdly subjective?

Also, some people are putting tags not related to the project just so they can get on trending. Some people put #tutorial in the title of games. This is only because it’s *nearly* impossible to get on trending without resharing
Lataliat
Scratcher
100+ posts

Is trending weirdly subjective?

Oh this forum is beautiful. I don't use the scratch search bar. I use the front page, and I also hop from person to person via comments and projects really. A long time ago, this is how it used to be. I was most active on Scratch in 2015 and 2016. The KINDS of projects that were popular then were different. I know this has multiple sources, but I don't think this trending idea helps. I personally am someone who repeatedly visits the Scratchers I am familiar with or friends with. I think it is kind of crazy how numbers-driven certain Scratchers have become.
kriblo
Scratcher
100+ posts

Is trending weirdly subjective?

Lataliat wrote:

Oh this forum is beautiful. I don't use the scratch search bar. I use the front page, and I also hop from person to person via comments and projects really. A long time ago, this is how it used to be. I was most active on Scratch in 2015 and 2016. The KINDS of projects that were popular then were different. I know this has multiple sources, but I don't think this trending idea helps. I personally am someone who repeatedly visits the Scratchers I am familiar with or friends with. I think it is kind of crazy how numbers-driven certain Scratchers have become.
I don't use the Explore page either, but it's obvious that many people do, as getting a project on Trending generates a lot of views. That's why I feel it needs to be relevant and fair. But, as I have previously suggested, another option is to remove this feature altogether.
Lataliat
Scratcher
100+ posts

Is trending weirdly subjective?

kriblo wrote:

I don't use the Explore page either, but it's obvious that many people do, as getting a project on Trending generates a lot of views. That's why I feel it needs to be relevant and fair. But, as I have previously suggested, another option is to remove this feature altogether.
Yeah that makes sense. Since I'm out of the loop, does it generate similar fame as to being featured? By the way I feel like removing this option could be a thing as long as they have something else. But I'm not really able to suggest alternatives.
colinmacc
Scratcher
1000+ posts

Is trending weirdly subjective?

Lataliat wrote:

kriblo wrote:

I don't use the Explore page either, but it's obvious that many people do, as getting a project on Trending generates a lot of views. That's why I feel it needs to be relevant and fair. But, as I have previously suggested, another option is to remove this feature altogether.
Yeah that makes sense. Since I'm out of the loop, does it generate similar fame as to being featured? By the way I feel like removing this option could be a thing as long as they have something else. But I'm not really able to suggest alternatives.

Yes, you get A LOT of views for being on the first 2 pages of trending.
Ihatr
Scratcher
1000+ posts

Is trending weirdly subjective?

I'm just down to removing trending altogether, because it has just turn into an all-out war for the most numbers. People resharing their games, uploading new games onto old games so they retain the numbers, it's just become a mess.

Nobody's going to agree with me on removing trending, but we need to do something to trending.
MrFluffyPenguins
Scratcher
1000+ posts

Is trending weirdly subjective?

Ihatr wrote:

I'm just down to removing trending altogether, because it has just turn into an all-out war for the most numbers. People resharing their games, uploading new games onto old games so they retain the numbers, it's just become a mess.

Nobody's going to agree with me on removing trending, but we need to do something to trending.
im down for removing trending
squig3
Scratcher
58 posts

Is trending weirdly subjective?

Lataliat wrote:

kriblo wrote:

I don't use the Explore page either, but it's obvious that many people do, as getting a project on Trending generates a lot of views. That's why I feel it needs to be relevant and fair. But, as I have previously suggested, another option is to remove this feature altogether.
Yeah that makes sense. Since I'm out of the loop, does it generate similar fame as to being featured? By the way I feel like removing this option could be a thing as long as they have something else. But I'm not really able to suggest alternatives.

I don't use the trending page so I'm also kind of out of the loop, but i've had a few projects get up to number 10-20 on trending since people will point it out in the comments. It does have a similar effect to being featured (i'm sure if you're number one or two it will be very similar to being featured). It's not going to yield as much recognition (the projects I have that have been featured got much more attention than the projects on trending, since the trending page moves quicker than the featured section). But it still is a substantial amount of fame. It is also far more accessible (or seems that way, since it's not curated by people) so that's why there's so much emphasis put on it by scratchers.

I'm definitely not opposed to removing the trending page altogether. As for what would replace it, I think a more in depth “What's Happening?” section would work. (at least for me)
I'd venture to guess about 90% of my traffic to scratch projects starts from seeing what the people I'm following are loving/favoriting/sharing, which is the part I really love about finding projects on scratch.
Here's a scenario that goes down quite often: Let's say I follow collinmacc, and see that he loved and favorited a project called “Cardinal Chains” by RokCoder. I check it out, and then I really enjoy the project. I will love and favorite it to tell my followers about it, and since I enjoyed that project, I check out some of RokCoder's other projects and enjoy those as well. I'll follow RokCoder not only because I want to be notified when he releases a new project, but also because he and I have a somewhat similar taste in projects, so projects he likes, I will probably like as well (and the cycle continues).

As of now though, the only way I decide whether I want to check out a project that someone else loved or shared is through the title, as that's all that appears in the What's Happening? section. Replacing the trending tab with a feed of followed scratchers' activity which display projects similar to how they are displayed on trending (title, thumbnail and author) would at least give me better chance to discover new projects / scratchers I'm interested in, and I have slight control over what I see by choosing who I follow.

Then again, that may make scratch too much like a social media site and not a place to share creative ideas through code.
RokCoder
Scratcher
1000+ posts

Is trending weirdly subjective?

Ihatr wrote:

I'm just down to removing trending altogether, because it has just turn into an all-out war for the most numbers. People resharing their games, uploading new games onto old games so they retain the numbers, it's just become a mess.

Nobody's going to agree with me on removing trending, but we need to do something to trending.
It wouldn't need removing if it felt like it was working correctly. My suggestion would be to only consider views (or loves or favourites) over the first x days since being originally shared (let's say 20 days) and count the number of views (or loves or favourites) between 4 days ago and the current date. On the first day of sharing you'd have one day's worth of views; on the fifth day of sharing, and each subsequent day up to twenty, you'd have five days of views. After twenty you'd have four then three then two then one day of views and then you'd be out of trending completely. This would allow projects to organically rise up and drift out of trending, would be based solely from the original share date and wouldn't allow any project to stay at the top for more than two weeks (if x is twenty days). To me that sounds like a fair system that wouldn't allow for any cheating. Maybe I'm missing something?
Maximouse
Scratcher
1000+ posts

Is trending weirdly subjective?

RokCoder wrote:

It wouldn't need removing if it felt like it was working correctly. My suggestion would be to only consider views (or loves or favourites) over the first x days since being originally shared (let's say 20 days) and count the number of views (or loves or favourites) between 4 days ago and the current date. On the first day of sharing you'd have one day's worth of views; on the fifth day of sharing, and each subsequent day up to twenty, you'd have five days of views. After twenty you'd have four then three then two then one day of views and then you'd be out of trending completely. This would allow projects to organically rise up and drift out of trending, would be based solely from the original share date and wouldn't allow any project to stay at the top for more than two weeks (if x is twenty days). To me that sounds like a fair system that wouldn't allow for any cheating. Maybe I'm missing something?
This would probably work, although it it was implemented Scratch would need to remember when the project was first shared – I don't think it currently does.
kriblo
Scratcher
100+ posts

Is trending weirdly subjective?

RokCoder wrote:

Ihatr wrote:

I'm just down to removing trending altogether, because it has just turn into an all-out war for the most numbers. People resharing their games, uploading new games onto old games so they retain the numbers, it's just become a mess.

Nobody's going to agree with me on removing trending, but we need to do something to trending.
It wouldn't need removing if it felt like it was working correctly. My suggestion would be to only consider views (or loves or favourites) over the first x days since being originally shared (let's say 20 days) and count the number of views (or loves or favourites) between 4 days ago and the current date. On the first day of sharing you'd have one day's worth of views; on the fifth day of sharing, and each subsequent day up to twenty, you'd have five days of views. After twenty you'd have four then three then two then one day of views and then you'd be out of trending completely. This would allow projects to organically rise up and drift out of trending, would be based solely from the original share date and wouldn't allow any project to stay at the top for more than two weeks (if x is twenty days). To me that sounds like a fair system that wouldn't allow for any cheating. Maybe I'm missing something?

I fully agree. Personally I don't see a problem with listing trending and popular projects, and it seems many Scratchers do want to know what other Scratchers view and appreciate. The approach you suggest seem reasonable. But if, for some reason, you wanted to prevent popular Scratchers and projects from getting even more views, it's better to remove the “Trending” and “Popular” pages, and have “Editorial” and “Random” in its place.
colinmacc
Scratcher
1000+ posts

Is trending weirdly subjective?

The way I see it, the trending and popular pages should be a good place to view the very best that Scratch has to offer, but if they are deliberately excluding the very best and the most popular projects then they don't really serve any purpose.

At the moment, it doesn't seem as if they do represent the very best of Scratch.
Ihatr
Scratcher
1000+ posts

Is trending weirdly subjective?

colinmacc wrote:

At the moment, it doesn't seem as if they do represent the very best of Scratch.
There isn't a single project on trending that I would say is unique. Almost every single thing on trending is some random platformer, some meme project, or some animation that could have been created in 5 minutes

What's so annoying is that the best Scratch projects never get on trending. Never. A good majority of them barely get noticed, and by the slim chance they do, they just get 50 hearts.

Scratch's project quality is rapidly decreasing because people know that they can get popular without having to put lots of effort into their projects. All Scratcher's care about is the numbers, and thus they push out something that doesn't deserve to be popular, and people appreciate them for it. So many good coders have focused on the numbers instead of quality, and thus their projects become something I don't enjoy anymore.

It's really sad to see what the trending page has done to Scratch.
MeIzAwezomeDede
Scratcher
1000+ posts

Is trending weirdly subjective?

Ihatr wrote:

There isn't a single project on trending that I would say is unique. Almost every single thing on trending is some random platformer, some meme project, or some animation that could have been created in 5 minutes
Me omw (on my way) to create generic platformer #289

I do agree though the amount of copy + paste projects is awful, I'd say about 5% of the projects on trending actually had effort put into them

Powered by DjangoBB