Currently, attempting to submit a run to a category with subcategories results in the subcategory fields being filled with the default information, regardless of which subcategory is currently visible. As an example using Bullet Heaven 2: https://www.speedrun.com/bh2

The follow set of categories are the default: https://prnt.sc/kmmt4q The following set of categories are then looked at/set to be visible: https://prnt.sc/kmmsue Hitting "Submit run" populates the submission with the defaults instead of the selected categories: https://prnt.sc/kmmsw5

I noticed this strange behavior when populating the Ape Escape Category Extensions board, of all things- several runs ended up in the wrong category at first, and I've seen a few submissions slip through from the wrong subcategories as well because you would expect to be submitting a run to the currently visible leaderboard, as with non-subcategorized categories.

Ask one of the site mods to roll back the board to the last backup: I've seen this done before.

Interesting- I had suspected it was an emulation setting difference, not that it could be a version difference. If I had (or could acquire) that version I'd run some timing comparisons between your version, kupo's version, Anthology on my machine, and Dosbox defaults to see if it's version or computer dependent.

edit: actually, is the version available on Steam the same as the anthology version?

I was looking into the runs here and noticed that Zen_Aku's game seemed to be running faster than gerudokupo's. So I decided to look into it to see if I was crazy, and if I wasn't, whether it meant a game was running slow or a game was running fast. Here are my findings: https://imgur.com/a/yuuw3oE

As shown, Zen_Aku's game is running faster, and from my own timings (not recorded) gerudokupo's game is running the same speed as mine. This obviously creates a problem where incorrect emulation is granting a sizable advantage- if the 18:38 was running the same speed as the 19:34, it would be roughly 21:00!

The simplest course of action would be to remove the offending runs that have the game running too fast. In the past when the game speed increase has been an accident, however, I've seen run timings adjusted to what they "should" have been to not invalidate them entirely.

@Tenka forum alerts aren't guaranteed, which is why I went here with a general inquiry (which morphed into a lot more specifics to try to explain the situation...) after trying to find a way to get into direct contact. Default DOSbox settings, for me, matched other footage of both DOSbox and non-emulated, leaving only that run as an outlier.

I'll try contacting through Twitter and also compile supporting documentation to post in the game forums. Thanks for the link @Liv !

For now, this is an academic inquiry as I'm hoping to be able to contact the mod in question. But in case that falls through, I'm wondering what the correct procedure would be for fixing the problems- would I need to request mod privileges myself to fix things? Normally I'd just drop a forum post in the game and ping on Twitter, but the mod has no linked Twitter with Twitch whispers off and a private Steam profile. Compounding the issue, it's a DOS game and I noted that the mods' runs are running 13% faster than default settings (confirmed on several machines and vs. other gameplay footage).

The game(s) are Math Rescue and Word Rescue, because I recently set up DOSbox to play them again and, naturally, ended up going for ILs in Math Rescue as it had an in-game timer. Which took me to the SRcom pages, with runs, and a poorly played time on a level with no tech 1 second faster than possible at default speed.

Assuming I can't get in contact and get mod privileges, what would be the correct etiquette? My thoughts would be to adjust the time on the runs to what the run would have been at normal speed (Word Rescue would add 8 seconds per minute, Math Rescue would add 9 seconds per minute, both timed to the frame on long, straight stretches) with a note similar to what was done for Embodiment of Scarlet Devil and its 62.5 (instead of 60) FPS runs rather than reject. Polishing up the categories and rules are a given.

tl;dr DOS game speedrunning can be a mess, sole "active" mod can't be contacted, has game running too fast (I hesitate to call it cheating as it's likely inadvertent), and only has a single category for Math Rescue due to thinking higher difficulties are impossible (they aren't), wondering about the "correct" course of action.

I see that there's currently only an "any%" category for the game, with no difficulty settings or 100%. I also noticed that the comment on episode 1 said that the higher difficulties don't work, which isn't correct- they just have a slightly different requirement. Anyway, proposed categories (assuming for all 3 episodes but I can't personally say for episodes 2 or 3):

Easy: Requires collecting any 10 numbers to open the door. Medium/Hard: Requires collecting all 20 numbers to open the door. The difference here is in Gruzzle behavior- Hard has them respawn and be "smarter", as well as having less slime available. 100%: Requires collecting all post-level bonuses. This requires not having "Skip problems" selected and collecting the numbers from 0-9 twice, getting all trucks correct, getting all problems correct, collecting all collectibles, and accessing all secret areas. Damage can be taken. Due to the forced math problems, a "100% collection" category that meets all of the above requirements without the math problems would probably be a better category- less RNG from the mid-level problems, no interruptions.

In summary: Any% Easy, Any% Medium/Hard, All Collectibles, and 100% would be the ideal category list for this game.

"Well at least in the US where I live almost everyone I know has stopped using the desktops. Statistics also show a fairly rapid decline in usage."

I'm sure that's partially due to smartphones taking off, as implied, but also in large part due to laptops reaching parity with the average desktop, with the added benefit of portability. When upgrading a computer, that's a real tangible benefit and current laptops exceed the desktops of yesteryear. At least for me, I see a few desktops around homes and those that have them often have a laptop as well. And a good laptop these days can run a PS2 emulator fairly well, and this is only going to get easier with further technological advances.

Dolphin has been getting pretty accurate lately, so I wouldn't be surprised if it became allowed at some point in the future.

PS2/XBox emulation is still quite far off, but GC is "close".

Guys, the link Hummeldon posted 6 months ago still works. It's literally 3 posts up from mine.

It's worth noting that most games that don't allow emulators are because the emulation isn't accurate enough, often giving an advantage to the emulator. This doesn't mean you can't do runs on an emulator, just that they can't be submitted to a leaderboard.

The rule used to be "top 10 or top 50% (whichever is smaller) needs a video" even here. But this was recently hanged due to a shift in community opinions on proof standards, which is fair.

Before speedrun.com everything was on a google doc, which had to be manually updated. As a result, a lot of runs were added after they were seen live before they were highlighted, and some of those runs never got highlighted because the runner intended to improve again soon, but that improvement never happened.

http://www.speedrunslive.com/races/result/#!/76693 39 hours, RTA, the only times I stepped away from the game were food/bathroom breaks. At the time of completion this was actually the longest finish time on any SRL race to date, though it's since been surpassed by a few other marathon races (GTA 100% trilogy had a 66 hour finish time, I recall, and the Yugioh Forbidden Memories 100% lasted 5 days with sleep breaks before SRL admins shut it down) but I think it's still the longest first place time. And it is a genuine first place- Zorasknight dropped out 33 hours in when it was clear I was going to finish first by a lot.

This is my current PB for the category, coming in about 10 hours faster. I'm going to knock it under 24 someday, but it's very hard to find the time to do a run :P

There's another option other than a capture card- if you don't mind the loss of quality, you can always point a camera at the screen and stream/record the camera.

Well, VC and disc were split into subcategories before there were complaints... I see no reason why that couldn't be re-done if there was demand for it.

0.3 seconds per loading screen spread across dozens of loading screens would add up pretty quickly. 25 levels + 5 bosses = 30 entry loading screens, 29 exit. That comes out to ~18 seconds lost to loading screens. Not much time, but enough to really matter at the top levels of play.

"Anybody that runs the game knows what the WR is, regardless of a leaderboard. "

This is simply not true, especially when the leaderboard doesn't reflect the actual record. Sure, some runners will know and pass on this information, but what about the people who don't interact with the community much? What if they searched for a video, and it's on Nicovideo and they only checked youtube? Twitch is awful for searching, and so many runs have their video there.

And then what about the people who want to start running a game, or started after the run was removed? How are they supposed to know, especially in smaller communities?

I know this is a double post, but another point occurred to me.

For the people who say that a time without a video is meaningless... what happens if a fraudulent copyright strike takes down the youtube upload? What if twitch glitches and loses the video on their servers? The video used to exist, just like with runs where the runner took down the video, and was verified as an actual run that took place and was correctly timed.

Should runs that are lost by no fault of the runner be invalidated as well? If not, why is having a time with a removed video any worse?

The Ape Escape speedrunning community has a Discord of its own. It's still a little small due to some historic overlap with other communities, but it's still there.

So, here's the invite: https://discord.gg/8CQxxBT

On the behalf of the AE speedrunning community, we hope to see you in here soon! :)

At the same time, though, there has been at least one case of site staff stepping in and overturning the community decision.

For both viewpoints.

That, to me, says that a concrete guideline for this should be implemented as a reasonable default that individual communities can decide to do differently... but when the community wants to leave a time up with no identifying information and someone on the staff sides with the guy who deleted an entire leaderboard I think that there's an issue.