If it catches on like this, everyone will have to follow suit. Just go every day and download your own models at irregular intervals. Maybe you can run a PHP script on Apache to do it automatically. You might find a way to use changing IP addresses. There are also programmes that record the keyboard and mouse. Recorder. You can replay the scripts, even repeat them. To have changing IPs, Tor-Browser might be a good alternative 
Scrolling through the models on the “Trending” tab I only see well designed and documented models by a large variety of very talented and well known creators.
5 Likes
Not if you sort by the last 7 days. The trending of all time is very slow to change so they wont show up there.
So what exactly is the issue then? Isn’t the system doing exactly what it should be doing?
Oh and I found your bot…
5 Likes
Lol, you only need to look as far as the #3 ranked model on trending from this week.
Nik00lo currently occupies 5 of the top 11 trending models with all of those models posted at the same time yesterday and all achieving 1:1 or higher likes to downloads.
Let’s look at the comments on their most popular model which has only a screenshot of a shoe from the slicer render. Nobody shares a photo of the print and note that the model doesn’t have any overhangs despite being praised for them.
Finally, following up on these commenters they are all empty accounts that only follow Nik00lo or their other point farming alt, Ziobanana.
Now let’s go look at Nik00lo’s most recent model uploaded 9 hours ago and already 10 comments. (They have already been accused of using a stolen model and photo for this one)
Do any of those comments look familiar?
1 Like
So we should all only ever sort by “All time”? The other ones will become useless if they are bloated with stuff that has fake download numbers. This is a quite clear issue in my opinion.
1 Like
Well it probably defaults to “All Time” for a reason?
You’re not wrong about this.
The thing is that there will always be people that try to game the system via scripts and alt accounts especially when a platform offers a monetary gain for doing it. Do you have a solution to eradicating it? If you do I’m sure YouTube would love to have a chat with you as they have yet to figure out how to solve this issue that plagues their platform. It’s just one of those things you really can’t avoid after a platform reaches a certain amount of popularity and it’s a near impossible thing to stay on top of as popularity grows.
4 Likes
If “all time” users are in the top positions, won’t their models be loaded more often because they appear first? Is it possible to oust them from these places at all?
Look at user accounts for activity, their distribution of activity during the day and what variety of content they engage with to establish a bot likelihood measure. For accounts that fall below minimum threshold simply don’t use their input when calculating rankings or points. If they cross the threshold later, add their input back in.
Look at models for statistical outliers in terms of gaining likes and downloads quickly or ratio of likes to downloads. 20% likes/downloads is high for a popular model, yet many abusers carry a like ratio above 100%. Flag them for review. Look at the users that like and download the same model, are they clumped in time? Do they browse and engage with other content or only the one user? The trending ranking is already a very good measure of potential bot abuse.
Look at originating click source. Are their fans linking from a forum or discord or something (likely not bot) vs are they all searching or direct linking to content.
Lastly, simply browse for obvious abusers. There’s not a ton of content and it’s not hard to see, a single part time employee could easily crack down on the obvious abuse going on right now.
So your solution is to manually review and delete bots? That’s not a solution, that’s just a bandaid. It doesn’t solve the bot problem as more bots are easily created using the original method. They can be created faster than it takes to review and delete them and they can be created around the clock.
4 Likes
Let’s walk through another example. This is PrintOnDemand, one of the more popular designers on Makerworld. All of their models start with a minimum of 42 likes and they carry a greater than 1:1 like to download ratio on models up until they start to surpass 60 downloads.
It doesn’t take a genius to see what’s going on. Their work gets traction quickly because they always start near the top of the rankings.
Here’s a model with 51 downloads and 47 likes, pretty high ratio. Let’s see who comments on it. torpid beaver loved it. Whose torpid beaver? Oh look, they’re a fan of PrintOnDemand and another very similar French designer, Remi’s Creation, and nothingtoloose.
Lets move on. Here’s another model with 32 downloads and 49 likes, a very popular set of cups. Again one comment, innocent herring thinks it’s cool. Another animal-named account. And who do they follow? Remi and Print On Demand.
Maybe a bit of a pattern sure. Let’s glance at another Print On Demand model. A dog fridge magnet with 22 downloads and 50 likes, more than 2:1 likes to downloads! Who comments on it? Another animal, feigned tapir thinks it’s cool. And who do they follow? Remi and Print.
If you look at Print On Demand’s followers, they have at least 60 more of these animal themed bots following them and helping them reach the top of the trending category. And all of their models are filled with comments of “cool” “I love it” from the same pool of bots.
1 Like
Not manually, automatically based on a bot-scoring threshold. I laid it all out already. But manual takedowns is the least they can do when users are being this obvious about it.
Maker World is still in open beta. I’m sure lots will probably change going forward. I’m also sure all hands are on deck dealing with copyright issues as that is a much bigger problem in my opinion.
4 Likes
You 2400 dowloads???
Bambu labs :This person has more than 2400 downloads,bot??? and has called people CRIMINAL.
This ought to be fun to follow from here on out. 
If the posts don’t get deleted.
1 Like
im not a bot, nor an offender. I demand you to apologize right now and delete me from your childish list.
that makes me a bot? or an offender?
I didnt say you are a robot, I said your download numbers are botted or not real…