Non-Mainstream Movies Statistics in Usenet Indexers: A Sample

[sbestran]

Hello,

As part of my research on Usenet indexers, I've compiled statistics to compare the availability of "cult" (non-mainstream) movies across 18 different indexers.

While I understand that this small sample collected in the past 6 months is as particular as any ARRs statistics that we find around the web, this analysis aims to help inform about which indexers are most likely to have a more diverse range of global cinema.

About the scope these tables:

The sample set data consists of around 550 movies: approximately 330 non-English language films from various countries, along with about 200 English-speaking films (some from the USA, but many from other countries like the UK, Canada, Australia, South Africa, etc.).
The main genres are Drama, Comedy, Historical and a few Documentaries. Notable genre exclusions are Horror, Action, Western and Children's movies.
The movies dates are (rounded) 15% from 2020s, 25% from 2010s, 25% from 2000s, 10% from 1990s, 5% from the 1980s, 10% from the 1970s and 10% from the 1960s.
I tried to put a bit from every country, specially well known movies or directors from different countries (Including the US).
All other information like languages and available qualities are shown in the tables, including in which priorities the Indexers were used (Yes, different priorities, but that is also data).
Radarr was set to search for the higher quality possible, with Remuxes over Disks, over Rips, over Web, over DVD/SD (Like in the table disposition).
Only movies available trough RadARR for these 18 in were used.

1st TABLE: General Availability of Non-Mainstream Movies per Indexer and Qualities


Some observations I made to myself after compiling this table:

- Even with very low priorities, NZB.su and DrunkenSlug had more available movies than other "famous" Indexers like Planet and Geek.
- Finder was a big surprise.
- Ninja had lower score than I imagined. (Although still OK for a low priority Indexer)
- The major part of Dog's releases were high quality ones.
- Planet results were also a surprise.
- The Cat is really dead for this content.
- While DVD and WEB qualities were not the goal, finding them in indexers is a point that at least it had that content, and that no other indexer had it in better quality

2nd TABLE: International Language Movies per Indexers


Some observations I made to myself after compiling this table:

- DrunkenSlug and NZB.su are, again, very good for non-English movies.
- Finder won against Dog and Geek for overall non-English movies.
- Dog went better than all the others for English language movies.
- DrunkenSlug, NZB.su and Finder went well in French movies
- I expected Scene to go better with German movies, specially at 2nd Priority. Maybe the quality preset made it not go well?
- Spanish Language films (from both Spain and Latin America) were most found in Finder, Geek, Slug, su and Dog
- The sample data is very small to make considerations for any other language


Final thoughts:

This data is a small snapshot of the availability and international language coverage across various indexers based on my own dataset. The results are specific to this sample, but they aim to inform people searching for non-mainstream movies at maximum quality inside Usenet Indexers about them. It was made thinking about people who could be using "cinephile" trackers like SecretCinema, KG or CinemaTIK or anyone who likes these types of movies and feel a bit "lost" when searching for them in the Usenet. It is a set of statistics that when I searched I could only find people asking for them with no answer, so I resolved to make this and share. It is the type of thing I would love to have had at the beginning of my Usenet voyage.

Draw your own conclusions from the data and use it as a comparison or starting point for your own research. If you have any questions about the context on these findings, please ask.

If you have your own stats or experience on this area to share, be welcome to do it!

[sigduwksnsksis9283]

awesome work. i also found drunk and dog to be decently good, dog being better in quality curated but having shitty site that goes down, i sometimes i wish i could just scrape their entire site and make my own database of it.

And drunk is my goto for most content as its free is enough for me as i casually watch at best.

[anon]

Thanks, this has indeed been asked more than once and will be helpful for the 99% who isn't in KG.

[sbestran]

awesome work. i also found drunk and dog to be decently good, dog being better in quality curated but having shitty site that goes down, i sometimes i wish i could just scrape their entire site and make my own database of it.

And drunk is my goto for most content as its free is enough for me as i casually watch at best.

Thanks,
Yes, Dog has a hold on me, it's hard to leave: I wish I could somehow help their server to be stronger.
Specially for movies it really feels like they curate their releases (that often are nowhere else to be found).
However, Sometimes it takes like 10~15 seconds to even answer an api call, making it become disabled in ARRs. It's needed to go there and re-enable the indexer manually testing/saving it's settings. Not optimal, a bit annoying, but I still want to keep them with high priority after these results.

Drunk also surprised me in these results, I believed It was being used for series, but it seems it's a must have for international cinema too.

[anon]

dog being better in quality curated but having shitty site that goes down, i sometimes i wish i could just scrape their entire site and make my own database of it.

Sometimes it takes like 10~15 seconds to even answer an api call

Sounds like you two could join forces regarding the scraping...

I did a lookup, and apparently both the main site and API server are hosted in Costa Rica, which was an interesting find. Given the idiosyncratic peering arrangements in Latin America, using a proxy in the region may help, if the problem lies in networking and not running on a toaster.

[sbestran]

Sounds like you two could join forces regarding the scraping...

I did a lookup, and apparently both the main site and API server are hosted in Costa Rica, which was an interesting find. Given the idiosyncratic peering arrangements in Latin America, using a proxy in the region may help, if the problem lies in networking and not running on a toaster.

I don't think it's a networking/peering problem, Anon.
Last month has been better and today for instance is a good day and they are answering in approximately 4 seconds, from which 2,2s were spent in Europe reaching the US and then 0,7 from there to Costa Rica plus 0,75s in that Country. (Using OpenVisualTraceRoute2.0.0).

A recent update to ProwlARR made it a bit more resilient to disabling indexers on their 1st timeout, which was great for Dog who in "dog days" can take these 15s to respond (besides the server being online).

It really feels like they are(/were?) hosting in some cheap or shared hosting that sometimes get overloaded maybe by the site/api maybe by whatever else is running in their server machine, and I don't know if they are in a financial position to make a move on that. I can imagine both prices and anonymity as reasons for this floating performance (issues for 2 months and back to work ok for some time).

But yes, I understand sigduwksnsksis9283 when he wants to scrap the site: some times you can't choose your nzb because dog was just out of the results during the search, and you need to search again, since their results make a big difference when it comes to non-mainstream or harder to find movies. And then we feel like we are using a thing that can just fail forever from night to day.

I just wish they succeed in improving their response time (4s, as today, is fine!) and that they keep a good backup of their catalog, because it is indeed a well done catalog when it comes to this content.

[anon]

TCP traceroute consistently gave me an RTT of ~200 ms on a 16-hop path, but a run of 1000 TCP pings had a packet loss of 40.2%

I think your theory makes the most sense, but surely they could do a donation drive or something? Having a site that's broken half the time sucks for everyone and alienates both current and potential donors.