Popular New Titles issues

[smart_fella255]

Don't look at the top ten on the front page, then....

yeah the top 10 are pretty trash as well. i wonder how they even get up their, it's mostly just fantasy/isekai slop

 

[Angry_Panda]

yeah the top 10 are pretty trash as well. i wonder how they even get up their, it's mostly just fantasy/isekai slop

My understanding is the criteria are recent title additions (last 30 days?) and 'reading count' (added to library). If that's wrong, someone please correct me.

So it's literally a representation of the current tastes of readers, with respect to new titles on the site. Which means there are a bunch of users reading that stuff.

 

[bondejeger]

My understanding is the criteria are recent title additions (last 30 days?) and 'reading count' (added to library). If that's wrong, someone please correct me.

So it's literally a representation of the current tastes of readers, with respect to new titles on the site. Which means there are a bunch of users reading that stuff.

It's a little more complicated than that. It's title entries from the last 30 days yes, and in theory it follows the follower count, but the count is changed with new chapters and several chapters added at the same time, and also I think by not having a chapter for 7 days, and how old the entry is I think. So usually groups that try to hit the list will try to release chapters around when titles drop off the list. Though it might also just change randomly some times, I don't always understand how it works I must admit.

A lot of the manga on the list mirrors what's popular on reddit, as that is the biggest manga community. But a lot of the manga that's not popular on reddit like isekai manga (reddit is basically rom coms), I think gets followers from twitter, tik tok, and other sources. A lot of times they're translated by for profit groups with their own readers.

I have a feeling that a lot of people who adds the manga from the list don't really read them though, so I'm not really certain if you miss out on much.

 

[TestAddress]

My understanding is the criteria are recent title additions (last 30 days?) and 'reading count' (added to library). If that's wrong, someone please correct me.

So it's literally a representation of the current tastes of readers, with respect to new titles on the site. Which means there are a bunch of users reading that stuff.

True, it is a stat based on views. I think the reactions are an indirect reflection because, a reaction means a view is made then went to the comments section to react.

 

[RogueKitsune]

It's a little more complicated than that.

Its not.

I don't always understand how it works I must admit.

Indeed...

Popular new titles is really simple. For all titles created in the last 30 days (rounded up to the next hour), filter out titles that don't have chapters and don't match users selected content ratings, then return the first 10 titles when sorted by follower count descending. That is all there is to it. And yes the average MD reader has really trash tastes and is horny as fuck.

Example API Request:

https://api.mangadex.org/manga?includes[]=cover_art&includes[]=artist&includes[]=author&order[followedCount]=desc&contentRating[]=safe&contentRating[]=suggestive&contentRating[]=erotica&contentRating[]=pornographic&hasAvailableChapters=true&createdAtSince=2024-08-24T04%3A00%3A00

note: "includes" tell the api to return some extra information that is not normally returned.

Anyway back on topic...
I don't really care if someone uses AI to help make some art to go with their story or uses AI to help make a story and draws art to go with it. Now if they are using AI for the whole thing, that's a problem.

DISCLAIMER: My statements are based on the assumption that whatever AI model is being used, it was trained on data that was ethically (i.e. relevant permissions, consent, rights, etc. were granted) sourced.

 

[Teasday]

Popular new titles is really simple.

Yep. 10 titles that were created in the last 30 days and have at least one chapter, sorted by followed count (plus taking into account content filter settings). I really wonder where people get these wild headcanons from and then confidently present them as facts, especially in a case like this where checking it is really simple.

 

[bondejeger]

Yep. 10 titles that were created in the last 30 days and have at least one chapter, sorted by followed count (plus taking into account content filter settings). I really wonder where people get these wild headcanons from and then confidently present them as facts, especially in a case like this where checking it is really simple.

Kind of interesting how manga like
https://mangadex.org/title/138654de...tagai-no-shoutai-wo-shiranai-danjo-no-hanashi
with significantly higher follower count at the time (I think it was 5000) got bypassed by manga with half the amount of followers if that's the case.
Yeah, checking it is really simple. You just see which manga hits the list when one drops off. And usually it's not the one with the highest follower count.

 

[Xnot]

Kind of interesting how manga like
https://mangadex.org/title/138654de...tagai-no-shoutai-wo-shiranai-danjo-no-hanashi
with significantly higher follower count at the time (I think it was 5000) got bypassed by manga with half the amount of followers if that's the case.
Yeah, checking it is really simple. You just see which manga hits the list when one drops off. And usually it's not the one with the highest follower count.

Everything you see on the site/API goes through multiple layers of cache and there are inconsistencies in the cache keys used between certain endpoints as well as when they get invalidated, so the follower count you see on titles is not always the same as the one the sorter sees.

 

[Teasday]

Kind of interesting how manga like
https://mangadex.org/title/138654de...tagai-no-shoutai-wo-shiranai-danjo-no-hanashi
with significantly higher follower count at the time (I think it was 5000) got bypassed by manga with half the amount of followers if that's the case.

?

I seriously doubt that actually happened. If you see a reproducible example of that happening, please report it because it would mean our API is buggy. As you can see from the URL, it's literally just using the exact same generic manga list endpoint that for example the search uses, not a special one where we have... maliciously changed the sorting order of a few entries in a list that on the frontpage starts from a random position and is constantly rotated anyway? What on earth would be the point of doing that?

edit: Oh right, the caching. Yeah that might explain it, especially at lower numbers. That's usually catching up daily, though.

Yeah, checking it is really simple. You just see which manga hits the list when one drops off. And usually it's not the one with the highest follower count.

Yeah, it's usually one of the lowest, because otherwise it would've been higher up on the list before. What do you mean?

 

[bondejeger]

maliciously changed the sorting order of a few entries in a list that on the frontpage starts from a random position and is constantly rotated anyway? What on earth would be the point of doing that?

The ones who know how it works are able to get their manga on the list, and the ones who don't, don't.

Yeah, it's usually one of the lowest, because otherwise it would've been higher up on the list before. What do you mean?

When a title drop off the list because it's older than 30 days, you don't always get a new one with the highest follower count the last 30 days. You usually get one with recently released chapters and lower follower count.

But yeah, this is getting off topic and I didn't mean to make a discussion out of it.

 

[Angry_Panda]

The ones who know how it works are able to get their manga on the list, and the ones who don't, don't.

Ah yes, the 'conspiracy theory'... uh... theory... of MD operations.

(no one mention the yacht)

 

[Teasday]

The ones who know how it works are able to get their manga on the list, and the ones who don't, don't.

What does this even mean...? It works by the follower count, knowing or not knowing that doesn't change anything.

When a title drop off the list because it's older than 30 days, you don't always get a new one with the highest follower count the last 30 days. You usually get one with recently released chapters and lower follower count.

You understand that when a title drops off from the top 10, then it's the 11th one that gets bumped to the top 10, right? Of course it's going to be one with a lower follower count, that's why it wasn't in the top 10 previously.

 

[bondejeger]

What does this even mean...? It works by the follower count, knowing or not knowing that doesn't change anything.


You understand that when a title drops off from the top 10, then it's the 11th one that gets bumped to the top 10, right? Of course it's going to be one with a lower follower count, that's why it wasn't in the top 10 previously.

I'm tired of discussing this with you, especially since you're discussing in bad faith.

 

[Teasday]

I'm tired of discussing this with you, especially since you're discussing in bad faith.

Right, just check out, admit nothing and accuse the other person of lying as a parting shot. Great.

 

[bondejeger]

Right, just check out, admit nothing and accuse the other person of lying as a parting shot. Great.

Why don't you check if the things I've been writing is correct or not when titles drop off the list in about four-five days? Apparently it doesn't help that I provide obvious examples. I don't understand what else you want me to do.

 

[Roler]

The ones who know how it works are able to get their manga on the list, and the ones who don't, don't.

When a title drop off the list because it's older than 30 days, you don't always get a new one with the highest follower count the last 30 days. You usually get one with recently released chapters and lower follower count.

But yeah, this is getting off topic and I didn't mean to make a discussion out of it.

It's sure nice to pull info from your ass.

It's easy to check because you can easily inspect all API requests the site makes by checking the network tab in your browser's dev tools.

Popular New Titles is not a pre-made list, it's a simple API request with the query like RogueKitsune sent.
Note: RogueKitsune also included erotica and pornographic, but that's not included unless you have porn enabled with other content ratings, or only erotica, because erotica was excluded in a frontend update if you have only the default content ratings enabled in order for it to be "less horny", it didn't help much tho...

So if you have only the default ratings enabled, the request URL for today would be this: https://api.mangadex.org/manga?includes[]=cover_art&includes[]=artist&includes[]=author&order[followedCount]=desc&contentRating[]=safe&contentRating[]=suggestive&hasAvailableChapters=true&createdAtSince=2024-08-24T22:00:00

However this result does not actually end at 10, as of writing this, there are a total of 275 titles in the result.
Here are the first 100: https://api.mangadex.org/manga?limit=100&offset=0&includes[]=cover_art&includes[]=artist&includes[]=author&order[followedCount]=desc&contentRating[]=safe&contentRating[]=suggestive&hasAvailableChapters=true&createdAtSince=2024-08-24T22:00:00

another 100: https://api.mangadex.org/manga?limit=100&offset=100&includes[]=cover_art&includes[]=artist&includes[]=author&order[followedCount]=desc&contentRating[]=safe&contentRating[]=suggestive&hasAvailableChapters=true&createdAtSince=2024-08-24T22:00:00

and the rest: https://api.mangadex.org/manga?limit=100&offset=200&includes[]=cover_art&includes[]=artist&includes[]=author&order[followedCount]=desc&contentRating[]=safe&contentRating[]=suggestive&hasAvailableChapters=true&createdAtSince=2024-08-24T22:00:00

and yes, MangaDex relies on caching in order to work, so you might see inconsistencies between some titles until the caches are cleared (and then again when the follow counts change).

 

[DavidianMillerian]

Why don't you check if the things I've been writing is correct or not when titles drop off the list in about four-five days? Apparently it doesn't help that I provide obvious examples. I don't understand what else you want me to do.

why don't you check if you're wrong?

 

[Teasday]

Why don't you check if the things I've been writing is correct or not when titles drop off the list in about four-five days? Apparently it doesn't help that I provide obvious examples. I don't understand what else you want me to do.

You haven't provided "obvious examples", you mentioned one title from 2 or 3 months ago for which I obviously can't check the validity of the claim anymore.

Someone else mentioned this:

I think their point is that its not the 11th title that usually takes the place in the top 10 when one there drops out. Saying its usually a title with lower follower count than the 11th that goes up there.

If that's what you mean, I really didn't get that impression at all, but it does make more sense than you being confused about the concept of No. 11 becoming No. 10.

In that case, my best guess for what you're seeing is that you think the title creation date is the same as the timestamp for the first chapter. Unfortunately it's not, and the title itself may be created a lot earlier than the first available chapter, which is a problem of fairness for how PNT works that we're aware of but do not have the ability to account for with the tools the frontend has. However, for the title you mentioned that's not the case since both were created 2024/8/8, so I'd still be leaning on follower count caching inconsistencies assuming you were actually correct about it. Inconsistencies in the couple thousands range aren't that uncommon, and I think that's usually around the natural range from No. 11 onwards, so it'd make sense.

 

[bondejeger]

You haven't provided "obvious examples", you mentioned one title from 2 or 3 months ago for which I obviously can't check the validity of the claim anymore.

Someone else mentioned this:

If that's what you mean, I really didn't get that impression at all, but it does make more sense than you being confused about the concept of No. 11 becoming No. 10.

In that case, my best guess for what you're seeing is that you think the title creation date is the same as the timestamp for the first chapter. Unfortunately it's not, and the title itself may be created a lot earlier than the first available chapter, which is a problem of fairness for how PNT works that we're aware of but do not have the ability to account for with the tools the frontend has. However, for the title you mentioned that's not the case since both were created 2024/8/8, so I'd still be leaning on follower count caching inconsistencies assuming you were actually correct about it.

It had a follower count of 5000 for a month, and never hit the list. It was higher than any other title that bypassed it. I don't see how this isn't obvious, even if you weren't there to watch it. You can't blame this on cache when it was like that for a month.

 

[Teasday]

It had a follower count of 5000 for a month, and never hit the list. It was higher than any other title that bypassed it. I don't see how this isn't obvious, even if you weren't there to watch it. You can't blame this on cache when it was like that for a month.

As Xnot explained, the inconsistency with the caching means that the number you see on the title entry and the number the manga list uses for ordering the entries can be different.

He just checked and it seems like the title page numbers are more accurate, so that can screw over the PNT order.

In other words, the PNT list works exactly as explained, just with incorrect numbers in some cases. We'll see if we can do anything about that.

wew I just hope we can get the view counts working some day and get more accurate metrics for this nonsense