Joined: 11/13/2006
Posts: 2908
Location: Northern California
One of the biggest things TASVideos has yet to implement is some sort of support for individual level runs. It's been requested many times, it's been something we've thought about many times, and in all of those many times we weren't really able to come up with a way of implementing them into our current methodology. You might be asking why we don't just change our methodology, to which I'd respond that it's so tightly integrated into the site that some of the principles and standards we've been working under for two decades might need to "die" to be able to change it.
That being said, I think ILs are way too important to just keep pushing back like this. We do want to support them, we have for a while, and it's finally time to come up with a way to do so.
There's a lot that needs to be discussed and refined here:
IL Submissions: Is it feasible to use the current submission system? Do we even want to do that? How would we need to change it to accommodate ILs?
IL Publication: How would encodes be handled? Would IL submitters even want encodes? Would it be too much work to encode every run that comes in, when ILs have the possibility to update several times a day in active communities?
IL Moderation: Would managing ILs be too much extra work for us as staff? Could they be handled by community members? How do we determine eligibility for IL moderation?
IL User Experience: How should ILs be displayed? How do we showcase them fairly in comparison to fullgame runs? How do we account for games that can only have ILs? How do we account for games that can't have ILs at all? Should every game have IL support by default even if there might not be distinct levels? Should ILs get submission and publication pages like fullgame runs? If not, how do we account for submission notes?
IL Verification: Given the nature of ILs, can we justify changing the verification standards for them assuming they haven't already been changed universally before IL implementation? That is, could we accept them using savestates or provided save files without needing verification movies? Just for archival/historical purposes, could we go as far as not even requiring input files, assuming we'd still heavily prioritize the ones that have them?
I'll be keeping track of ideas here, because it's a good way to end the OP without me having to fumble about for a sendoff.
Ideas will be tagged using the following identifiers:
(no tag) - The idea is new or has not received enough discussion for a tag yet.
(STAFF) - The idea was discussed internally by staff and is in need of community discussion.
(COMMUNITY) - The idea was discussed here by the community and is in need of internal discussion.
(APPROVED) - The idea has been discussed thoroughly by all parties and will be implemented if possible.
(REJECTED) - Either the staff or the community reached a consensus that this idea should not be implemented, or this idea directly conflicts with an approved idea.
As of the initial posting of this thread, most of the listed ideas are mine, with a couple of them having been provided/expanded upon by staff, though we're still too early on in discussion to have come to a consensus on anything. Ideas are not necessarily mutually exclusive and can work together unless they are directly opposed.
IL Submissions:IL Publication:
Encodes are not automatically provided to approved ILs, and are not required for approval as long as an input file is provided. Authors can opt in to having encodes made for them.
IL encodes made by us would go to a dedicated TASVideosChannel for ILs.
TASVideos may make backup encodes upon request, prioritizing author/community encodes as long as they are up.
IL Moderation:
Any staff member can "judge" and approve ILs.
There is no grace period for ILs, they can be approved as soon as they're submitted.
Community moderation is allowed for ILs: Each game can be assigned moderators, similar to SRC's moderation system, which would allow them to moderate ILs.
Joined: 3/10/2014
Posts: 16
Location: 🇺🇸 United States
> could we go as far as not even requiring input files,
Imo that would be a step too far, without input files, any preservation value is highly diminished I think.
Opting into encodes sounds decent but I would probably want to add a 2-4 week minimum of an IL to be unchanged before an encode is "triggered/enabled" to attempt to prevent cases of repetitive short-term obsoletion within reason.
One other aspect of this regarding quality, judging etc. is that some IL's are frankly.... uninteresting/trivially simplistic in TAS. the "filler episodes" of a fullgame run so to speak, that highlighting them individually seems somewhat silly. Playing a song is interesting, playing 1 chord is uhh... kinda mundane.
Another aspect is that IL's can end up having MANY more categories than fullgame runs, because a full run of a category might just simply be impossible to completing the entire game, but beating 1 level might be possible, and serves as an interesting puzzle to accomplish using the mechanics of the game. Categorization could end up tricky as you have each IL played with different Goals, using each individual character etc. the combinatorics quickly get really large in terms of displaying things in a way that makes any sort of coherent sense to an outsider.
Those are some of my initial impressions for now.
Joined: 10/12/2011
Posts: 6650
Location: The land down under.
I think personally there's more of a fine line to the ILs.
Some ILs are more destined for the Playground and some that are actually Publication worthy.
What I mean by that is take a look at games with IL clocks, say for example Crash Bandicoot: Warped.
This game has 30 levels, all with their own time trial scene-- split by region.
There's tricks that the TAS doesn't do, cause it's not beneficial for RTA compared to IL movement.
The problem with something like Crash, is that these levels all have standouts, but all are short.
This might be considered as "spam", and probably a way to deal with it in a sense is to have a major community IL mega file.
Otherwise it should be something more Playground worthy.
This extends to Crash Team Racing as well.
Tricks that are done without items is absolutely maddening, but the IL community there has been insane, but they've never considered submitting anything.
At least not yet, but it also has the better usage of actually having a Time Attack mode in the game making changing levels easy.
The second one is something we all know that I love bringing in the Space Ninja who stabs himself.
Well many newcomers don't know this, but there's fiveseparatesubmissionschallengingrules.
Maybe Tekken also falls into the "make it a major community compilation file", but the issue for Tekken compared to Crash is the scene is dead.
Tekken also almost falls into the "everyone is slower than Yoshimitsu, it might as well just be a Playground" but I tried a precedent, and it failed.
So honestly I don't know.
And yes I know it's based on the old rulesets, I'm still saying "I don't know" cause they would've been grabbed back when Playground was introduced.
As for Publications even that would have a line, since some people who make these ILs like putting in the most random ass music for whatever reason.
So even there's something to consider to point out in regards to if something is "Accepted".
I also agree that we actually need the files, you can get away with a lot of bullshit that people handwave over.
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account.Strong for yourself and also others.
The biggest problem with individual level TASes in my eyes is the initialization state. The initialization state needs to be properly accounted for, and not all games are as consistent on initialization with this regard... so you'll have to pay attention to that.
Admin Assistant, Senior Judge, Experienced player
(621)
🇨🇱 Chile
Joined: 7/8/2021
Posts: 188
Location: 🇨🇱 Chile
KungFuFurby wrote:
The biggest problem with individual level TASes in my eyes is the initialization state. The initialization state needs to be properly accounted for, and not all games are as consistent on initialization with this regard... so you'll have to pay attention to that.
There's a much larger conversation to be had about SaveRAM since it interacts with full game runs as well, and I want to reach a solution that reaches a good compromise between allowing people to make movies in the context they would be most comfortable with and making sure runtime environment is as reliable as possible. It is, in my opinion, the single largest issue to solve for our current methodologies, and I wonder if any attempt to make it work will allow our system to survive.
GuanlongX wrote:
auuuugh my face when games are creative and cool and don't fit arbitrary rulesets
The biggest problem with individual level TASes in my eyes is the initialization state. The initialization state needs to be properly accounted for, and not all games are as consistent on initialization with this regard... so you'll have to pay attention to that.
Would it be fair to assume the games that immediately benefit the most from individual level movie support tend to be the ones where it's possible to do the level selection in question from power up conveniently (with proper unlocking done via SRAM), ex: racing, fighting, etc?
Some of this might be controversial.
After completing Super Mario Bros. once, you have the option to choose a world, but not a stage within that world. The game state on starting a game this way is always the same regardless of how the game was completed (... mostly). Should ILs for Super Mario Bros. be worlds selected from the title screen, so each movie can start from a savestate (full game run verified) + soft reset allowing for traditional TAS timing as well as RTA and IGT? I think most people would consider an "individual level" to be a stage within the world, but then time could be affected by whether you start as small Mario, big Mario, or fire flower Mario, and also (depending on the timing method) what score you have at the start, on top of not having a clean "time from power on" or "time from soft reset" TAS time.
After completing Sonic 3 & Knuckles once, you have the option to choose a zone, but not an act within that zone. However, the game state on starting a game this way is dependent on the number of lives, continues, and emeralds earned while beating the game. The game also has a level select cheat that does allow you to select individual acts, and the game state on starting a game this way is always the same (... again, mostly). Should ILs for Sonic 3 & Knuckles be acts selected from the level select menu, so each movie can start from a hard reset allowing for traditional TAS timing as well as RTA and IGT? Would each character and super form have its own subcategory within this (potentially 300 in total)?
Is making Mario per-world using its "legit" level select feature but Sonic per-act ignoring its "legit" level select feature in favour of its level select cheat inconsistent? In both cases it's the smallest unit with consistent game state on starting, but that's not how it's perceived by the player. Should Sonic's ILs be limited to zones, but still selected from the level select cheat? If not, Lava Reef 2 and Death Egg 2 are each split into two levels internally (and on the level select menu) - do they count as two separate levels for the IL leaderboard?
Alex Kidd in Miracle World is a linear set of levels, separated by a map screen. But there's no level select, and the way you play a level can vary wildly depending on what items and how much money you brought forward from the previous level. Does it even make sense to have ILs for Alex Kidd in Miracle World if there's no consistent state in which you can start a level, and a level could be improved by playing the previous levels differently?
Going through points in the original post:
Samsara wrote:
IL Submissions: Is it feasible to use the current submission system? Do we even want to do that? How would we need to change it to accommodate ILs?
IL submissions should probably be a separate queue at minimum - there's a big difference in terms of movie count and length. The UI for the submitter seems as though it would be fine - just add a dropdown for full game/individual level and a textbox for the name of the level (or maybe a second dropdown?).
Samsara wrote:
IL Publication: How would encodes be handled? Would IL submitters even want encodes? Would it be too much work to encode every run that comes in, when ILs have the possibility to update several times a day in active communities?
[...]
Encodes are not automatically provided to approved ILs, and are not required for approval as long as an input file is provided. Authors can opt in to having encodes made for them.
IL encodes made by us would go to a dedicated TASVideosChannel for ILs.
TASVideos may make backup encodes upon request, prioritizing author/community encodes as long as they are up.
Given that it's already usually at least a few weeks for a full game run between submission and encoding, if an IL is updated several times a day then it'll be the most recent one that gets an encode a few weeks after it settles. I don't see any reason why this would need to be any different.
I agree with a separate channel - again, there's a big difference in terms of movie count and length. Perhaps if a game has recently had a large number of new IL runs the main channel could show a "compilation" of the ILs of that game (and use it as an excuse to advertise the second channel).
Samsara wrote:
IL Moderation: Would managing ILs be too much extra work for us as staff? Could they be handled by community members? How do we determine eligibility for IL moderation?
[...]
Any staff member can "judge" and approve ILs.
There is no grace period for ILs, they can be approved as soon as they're submitted.
Community moderation is allowed for ILs: Each game can be assigned moderators, similar to SRC's moderation system, which would allow them to moderate ILs.
I don't really have any answers or suggestions for this myself, but I feel like all of these can apply equally to full game runs as well?
Samsara wrote:
IL User Experience: How should ILs be displayed? How do we showcase them fairly in comparison to fullgame runs? How do we account for games that can only have ILs? How do we account for games that can't have ILs at all? Should every game have IL support by default even if there might not be distinct levels? Should ILs get submission and publication pages like fullgame runs? If not, how do we account for submission notes?
[...]
Any game with some way of identifying levels would get an IL table.
I like the mockup. It could be in an "IL" tab on the game page alongside other options.
ILs should definitely get submission and publication pages too. Full game submission notes should have the option to "include" an IL's notes if a full game run uses the IL.
I disagree quite strongly with the last point - I don't think it's enough to merely be able to identify levels, I think there needs to be a well-defined and consistent state for starting the level where performance in previous levels cannot affect the completion time. In other words, not only must there be levels, but they must also be individual.
Samsara wrote:
IL Verification: Given the nature of ILs, can we justify changing the verification standards for them assuming they haven't already been changed universally before IL implementation? That is, could we accept them using savestates or provided save files without needing verification movies? Just for archival/historical purposes, could we go as far as not even requiring input files, assuming we'd still heavily prioritize the ones that have them?
[...]
ILs can be approved with only a video, but will be marked as Unverified on the list. Unverified runs are hidden by default.
ILs are not required to have verification movies.
ILs are allowed to start from savestates.
Lowering verification requirements for ILs is equivalent to saying "we don't care about ILs" and I don't think that's the right way to go. Userfiles have always supported ILs without verification, just as they've always supported full game runs or partial runs without verification. If ILs are going to be promoted to publication, they need to have the requirements to go along with it.
After completing Super Mario Bros. once, you have the option to choose a world, but not a stage within that world. The game state on starting a game this way is always the same regardless of how the game was completed (... mostly).
What isn't the same? Could any part leading up to starting a world have any affect on the actual run? e.g. RNG (idk if that's a thing in SMB, but it could be in other games.)
DigitalDuck wrote:
Should ILs for Super Mario Bros. be worlds selected from the title screen, so each movie can start from a savestate (full game run verified) + soft reset allowing for traditional TAS timing as well as RTA and IGT? I think most people would consider an "individual level" to be a stage within the world, but then time could be affected by whether you start as small Mario, big Mario, or fire flower Mario, and also (depending on the timing method) what score you have at the start, on top of not having a clean "time from power on" or "time from soft reset" TAS time.
The starting state of the game, whether it be at the start of a world or a stage, definitely matters. I think a verification movie is a great approach for handling these questions. It not only does the regular job of verifying the lead up to the actual run is legit, it also sets the stage for what advantages/disadvantages the run receives at the start. For the sake of comparing which run is the "best", one would need to decide what's fair to compare in the first place. As in, is starting a stage with the fire flower just a speed saving technique, or is it a different category of run altogether from runs that start as small mario.
I think this is one of the major issues of this topic to be discussed. Personally, I think the lead up to the start of an IL matters and should be cataloged in some form. As such, it'd make sense to me if there was a different category for each kind of start, depending on what makes sense for the game. Still, there might be minute differences between verification movies for the same category that might lead to real differences in the actual run. I'm not sure how best to handle that. I don't want to say that everyone should be forced to use an prearranged verification movie; that would have its own set of issues and would likely stifle improvements. Maybe we just accept that verification movies can offer advantages and they become a more integral part of a TAS submission.
As for starting from reset or level-start, if we're using a verification movie, I don't think there's any need to start from reset. Alternative timing methods are something we're already working on implementing so this is less of a concern for IL discussions.
DigitalDuck wrote:
do they count as two separate levels for the IL leaderboard?
If an IL leaderboard becomes a thing, I don't think it'll be quite the same as what RTA runners may be used to, and that's okay. Fundamentally, TASing isn't always strictly comparable to human play. How a TAS is compared to another TAS may realistically be different to how an RTA run is compared to another RTA run.
Variations in the game's state may just be accepted as what it is to play a game as a human, since they are often not controllable. However, TASing offers a much much higher degree of control and precision. As such there are many more variables to consider when trying to compare one TAS to another TAS. If a run is trivialized down to just how fast the objective is completed, then it's simple to compare runs against each other. But if we care about how the objective is completed, then the comparison is harder to perform. It may not be something that can be simply displayed as a list of times.
IL Submissions: Is it feasible to use the current submission system? Do we even want to do that? How would we need to change it to accommodate ILs?
I feel that it would probably become overwhelming very quickly to use our current submission system. We definitely will need to outsource a bit and have something more built into game pages.
IL Publication: How would encodes be handled? Would IL submitters even want encodes? Would it be too much work to encode every run that comes in, when ILs have the possibility to update several times a day in active communities?
I think it's going to be game dependent. Some games might have ILs repeatedly update like you said, but others are likely going to treat their tables similar to full game publications currently. In any case, I feel an "Encode Requested" tag or flag authors can place on their own movies would have value.
IL Moderation: Would managing ILs be too much extra work for us as staff? Could they be handled by community members? How do we determine eligibility for IL moderation?
We're already worried about the scalability of how we handle fullgame runs, let alone ILs. Probably should be handled by communities to some extent, but obviously not every game supports a full community.
I have a feeling the answer to that last question is going to be "We just try some people and we get a better idea of what to look for over time."
IL User Experience: How should ILs be displayed? How do we showcase them fairly in comparison to fullgame runs? How do we account for games that can only have ILs? How do we account for games that can't have ILs at all? Should every game have IL support by default even if there might not be distinct levels? Should ILs get submission and publication pages like fullgame runs? If not, how do we account for submission notes?
I think games should have an "add IL table" option tied to a permission. If there's demand for IL support for a given game, it probably should be given. Dunno about pages, specifically.
IL Verification: Given the nature of ILs, can we justify changing the verification standards for them assuming they haven't already been changed universally before IL implementation? That is, could we accept them using savestates or provided save files without needing verification movies? Just for archival/historical purposes, could we go as far as not even requiring input files, assuming we'd still heavily prioritize the ones that have them?[/list]
I think in general we should move to more of a graded verification system as opposed to pass/fail, but that goes a bit outside the scope of this topic.
As for not requiring input files, I feel like I would want to hear from any communities who would like to use our services and see if that's something they'd want.
IL Publication:
Encodes are not automatically provided to approved ILs, and are not required for approval as long as an input file is provided. Authors can opt in to having encodes made for them.
IL encodes made by us would go to a dedicated TASVideosChannel for ILs.
TASVideos may make backup encodes upon request, prioritizing author/community encodes as long as they are up.
I agree with these largely. If we can, I would love to also see a compilation of the community's favorite ILs every month uploaded to the main channel, provided there's encodes.
IL Moderation:
Any staff member can "judge" and approve ILs.
There is no grace period for ILs, they can be approved as soon as they're submitted.
Community moderation is allowed for ILs: Each game can be assigned moderators, similar to SRC's moderation system, which would allow them to moderate ILs.
Any game with some way of identifying levels would get an IL table.
I'm thinking it might be better to use tab functionality for differing goals to prevent users from being overwhelmed. Maybe characters too by allowing for nested tabs. I think whatever we go with, it should be decently customizable per game page honestly.
I definitely agree that a full IL table should not be required
IL Verification:
ILs can be approved with only a video, but will be marked as Unverified on the list. Unverified runs are hidden by default.
ILs are not required to have verification movies.
ILs are allowed to start from savestates.
My thoughts on verification from before still apply.
[16:36:31] <Mothrayas> I have to say this argument about robot drug usage is a lot more fun than whatever else we have been doing in the past two+ hours
[16:08:10] <BenLubar> a TAS is just the limit of a segmented speedrun as the segment length approaches zero
This question is foundational! I personally see two paths of resolving it:
TASVideos hosts IL TASes like full-game TASes, with some remarks, of course.
TASVideos doesn't host IL TASes. It may be just an sharing platform. Like, a giant infinite-scrolling board with submission metadata for each entry everyone is free to use.
Is it feasible to use the current submission system? How would we need to change it to accommodate ILs?
I think the current submission system is feasible. However, I would propose several changes:
Introduce a separate entry ("Newest IL submissions") and queue for ILs, like it's done now for regular submissions.
We already have plenty of regular TAS submissions, some which end up being unseen, unless you regularly watch through the whole list. Having them mixed up with ILs will definitely clutter things up.
Introduce a separate, sortable field for levels/stages/chapters/worlds being TASed.
I think, "branches" field is already loaded enough. Especially, when people decide to put everything in it instead of, excuse me, "Game objectives" from submission pre-filled comments.
2. IL Publication:
How would encodes be handled? Would IL submitters even want encodes?
The main reason why people in general even make TASes is for popularisation. Hence, I would wrap it all into the question: "In which way should this popularization be done?" Here is my vision, although it may end up as a hybrid in some way.
Popularise the submitter's video.
"I have my own YT channel. I recorded a video of my IL TAS. I want more views, likes and comments" Sounds fair? I guess so. But in which way can this be achieved? Here I don't have any proposals, at least yet. I'll just share what I have.
I reviewed the analytics of my videos placed as submission encodes and the vast majority of them don't gain even 100 views from "tasvideos.org" + "discord.com". In case of playlists, the results are the same for the first video and fewer views for the later ones. This is not a kind of result I expect. 2k views and 15 comments in a published video looks much better in comparison.
Upload an encode to the somewhere already mentioned "TASVideosChannel for ILs" YouTube channel.
I personally don't mind if my IL video is published through some automatic pipeline, without human effort in VitualDub/AVISynth.
There might be a proxy storage which consumes a video file, that I as a TASer provide, and does some actions, like merging it with a 2-sec TASVideos.org intro video and uploading it on YT.
In any case, a video of an accepted IL should match a list of criteria, like raw footage and audio from start-up, no custom music, no user channel intro and so on.
3. IL Moderation:
Would managing ILs be too much extra work for us as staff? Could they be handled by community members? How do we determine eligibility for IL moderation?
It will certainly be time-consuming for staff, but how much - that's a question.
Community members may handle this, because an IL is certainly not a material for deep investigations. I would propose the following flow as a bare minimum:
A reviewer (or some other role) describes its proposal. Judge reviews it and basically says "Yes" or "No". Also, some timeouts might be introduced. Like, if a submission is proposed to be rejected, it's not necessary to wait for a judge to reach it in a pile of other submissions.
(I can share this draw.io diagram if needed)
4. IL User Experience:
How should ILs be displayed? How do we showcase them fairly in comparison to fullgame runs? How do we account for games that can only have ILs? How do we account for games that can't have ILs at all? Should every game have IL support by default even if there might not be distinct levels? Should ILs get submission and publication pages like fullgame runs? If not, how do we account for submission notes?
I mentioned the separate submissions and the board path above.
For publications path, I'd propose to mix them with full game TASes, but display them as a single game entry with some default description and a special tag or listing of ILs. Clicking on the entry would open the game page, like it is now with publications, users files and so on, with available ILs listed (maybe in a similar list with tabs as publications are). If a full TAS is present, it should be shown instead, but the ILs should be marked or mentioned somewhere there.
5. IL Verification:
Given the nature of ILs, can we justify changing the verification standards for them assuming they haven't already been changed universally before IL implementation? That is, could we accept them using savestates or provided save files without needing verification movies? Just for archival/historical purposes, could we go as far as not even requiring input files, assuming we'd still heavily prioritize the ones that have them?
Verification might become a necessary step for judgment, like a verdict proposal step I mentioned in the flow chart.
The question of omitting verification movies keeps raising periodically. I would say yes, we may accept them.
No, I'm against the absence of input files. We need to have some legitimate way of TAS verification, not just the provided video.
Also, I'd like to bring up the question of publishing unfinished TASes on request. A TASer may drop its progress due to various reasons. Instead of keeping a userfile somewhere, which may last untouched forever, I think it makes sense to let such works be published. At least as a IL TAS, although it might deserve to be treated as a regular TAS, depending on how much progress has been made.
"IL" should be something that's selectable. A level on the overworld - yes. A level within a world in a game which has selectable worlds - no. Submitting-publishing something that's not a full game TAS is already a good indulgence! Like, "C'mon! You don't feel like completing even a designed chunked of levels?"
coolman wrote:
I might be stupid on this, but most of time, wouldn't a full game TAS already have the best ILs? Since it need to beat or equals every existing records?
You're right, but a full game TAS might be constrained with its conditions. An IL might use all collected items/weapons for the sake of finishing a level faster, while a full game TAS has to keep them to be used later
TASing is like making a film: only the best takes are shown in the premier.
https://xkcd.com/3246/
Joined: 1/9/2023
Posts: 48
Location: Quebec Province
I might be stupid on this, but most of time, wouldn't a full game TAS already have the best ILs? Since it need to beat or equals every existing records?
Admin Assistant, Senior Judge, Experienced player
(621)
🇨🇱 Chile
Joined: 7/8/2021
Posts: 188
Location: 🇨🇱 Chile
coolman wrote:
I might be stupid on this, but most of time, wouldn't a full game TAS already have the best ILs? Since it need to beat or equals every existing records?
There's plenty of circumstances in which the best time for a level is not actually doable in a full game run. In Mario Kart Wii, for example, you cannot change character or vehicle between courses in a Grand Prix, so you choose the one that gives you the best sum time, meaning some courses will not be played with their optimal loadout. Sometimes a specific level may require specific RNG of framerules to get its best possible time that are not feasible or optimal when played in a full game setting.
GuanlongX wrote:
auuuugh my face when games are creative and cool and don't fit arbitrary rulesets
I might be stupid on this, but most of time, wouldn't a full game TAS already have the best ILs? Since it need to beat or equals every existing records?
Full game TASes require more commitment, so they can come out less frequently than ILs for plenty of games. There have been many cases where a full game TAS becomes outdated due to better ILs that came out later.
Additionally, not all games would even have all the levels in full game TASes, there can be extra levels in a menu after beating the game, there could be branching paths that aren't really covered by a fullgame run like Umihara Kawase, or a number of other reasons.
[16:36:31] <Mothrayas> I have to say this argument about robot drug usage is a lot more fun than whatever else we have been doing in the past two+ hours
[16:08:10] <BenLubar> a TAS is just the limit of a segmented speedrun as the segment length approaches zero
After completing Super Mario Bros. once, you have the option to choose a world, but not a stage within that world. The game state on starting a game this way is always the same regardless of how the game was completed (... mostly).
What isn't the same? Could any part leading up to starting a world have any affect on the actual run? e.g. RNG (idk if that's a thing in SMB, but it could be in other games.)
Setups that create things like ACE can change memory such that the state is different. For example, after finishing the recently submitted ACE TAS the game will always trigger ACE whenever a powerup appears, even though you've started a new game, which doesn't happen normally. At least, that's how I understand it.
Of course, in most circumstances that doesn't happen, so there is a "normal" state for worlds to start in, and that's what matters.
Bigbass wrote:
As for starting from reset or level-start, if we're using a verification movie, I don't think there's any need to start from reset. Alternative timing methods are something we're already working on implementing so this is less of a concern for IL discussions.
I agree; however, if there's a way to quickly choose a level from reset (e.g. via a level select code), it would prevent the need for a verification movie and I think that should be preferred.
Joined: 3/10/2014
Posts: 16
Location: 🇺🇸 United States
coolman wrote:
I might be stupid on this, but most of time, wouldn't a full game TAS already have the best ILs? Since it need to beat or equals every existing records?
Dustforce has several deathwarps which are used in IL (because the IL timer stops while dead and during fadeout/fadein/respawn) but which are slower in real-time because the time saved by the warp is smaller than the time the death takes. Thats just one example of a tradeoff that can occur with a "game timer" vs. RTA timing method, many IL's prioritize the former, while fullgame's tend to lean towards the latter as the accepted timing method.
Dustforce also has to account for both what is optimal inside the IL, but you ALSO use that character for the following overworld movement to reaching the next IL's entrance door. In several cases this means using a suboptimal IL character to achieve an overall faster IL+Overworld combined segment, saving time overall. And in several cases, the determining factor ends up being that a character is faster to menu select vs another character, such that the menuing-location becomes the determinant factor. an IL would avoid all of these conflicting and competing factors.
Is it feasible to use the current submission system? Do we even want to do that? How would we need to change it to accommodate ILs?
I have never submitted a movie, but I have read the submission rules (a long time ago), and just checked the submissions form. Nothing stands out to me as needing to change for ILs. Perhaps the stated goal could be expanded to provide a choice between "full game" vs. "individual level", with the IL option enabling a new field to specify the level.
Samsara wrote:
ILs have the possibility to update several times a day
If TASvideos is to be curating TASes, I do not think it makes sense to accept multiple submissions a day. If I remember correctly, one of the submission rules is that the TAS should be optimized. A run that is still being actively worked on is not ready for submission, IMO.
Samsara wrote:
could we accept them using savestates or provided save files without needing verification movies? Just for archival/historical purposes, could we go as far as not even requiring input files, assuming we'd still heavily prioritize the ones that have them?
I am in favor of allowing the use of save files without a verification movie, but allowing a movie to start from a savestate seems more problematic. The easier it is for users to download and play a movie themselves, the better; and using a savestate would demand the user download the specific emulator version that the movie was made on, which is (sometimes) not required without savestates. I suspect allowing movies to start from a savestate would also result in some setup being hidden, being done before the savestate, which can be frustrating both as a casual viewer and when working with the movie file.
In regards to ILs potentially having different starting conditions, I don't think TASvideos can prescribe the starting conditions that are allowed. There are situations where a particular starting condition is not optimal for the goal "fastest completion of this level" but still has entertainment value and can support competition within that category. There needs to be support for arbitrary categories within an IL even when a majority of the game's levels do not benefit from that. Some categories might exist in only a single level. This shouldn't be a problem at all for playarounds. When deciding which run is "fastest", I would think that in general you could determine if two runs are comparable by looking at the stated goals. For example "using X character", "starting with no items", "no glitches", etc. But there will probably be game-specific situations that require judgment, likely from a member of that game's community.
[*]IL Submissions: Is it feasible to use the current submission system? Do we even want to do that? How would we need to change it to accommodate ILs?
I would concur with what DigitalDuck, Memory and Dimon have said in the thread - ILs need to be handled through separate queues. Throwing every full-game submission on the site into the same assembly line is enough of a challenge to deal with as it stands. What does get more complicated is whether there should be one big IL queue, IL queues broken down per game, or both. I'm not completely sure of the answer myself, particularly as it really depends on what we end up deciding on regarding IL moderation.
Riokaii wrote:
One other aspect of this regarding quality, judging etc. is that some IL's are frankly.... uninteresting/trivially simplistic in TAS. the "filler episodes" of a fullgame run so to speak, that highlighting them individually seems somewhat silly. Playing a song is interesting, playing 1 chord is uhh... kinda mundane.
To me, I see a full-game run as more akin to an album or soundtrack, and an IL as more akin to a song on that album. And while it's true that some ILs in some games are going to be less interesting or easier to TAS...so are some entire games! I don't think we should try and set rules and standards about how interesting or difficult an IL needs to be to warrant publication - we already removed our entertainment and triviality requirements for that reason. We should instead try to make things as accessible if not more so.
I have similar feelings about Spike's suggestion that some ILs are deserving of publications and some ILs are more deserving of Playground, but I also want to address the suggestion of a "mega IL community submission" there. While in theory I believe this is a good idea - and one that can obsolete the need for per-level submissions - I think there are several challenges that make it more of a hindrance: IL compilations are bound to get updated far less often than ILs themselves (if at all), some games work better under a compilation framework than others, everyone in the community needs to be happy having their names together on the same product, etc.
Samsara wrote:
[*]IL Publication: How would encodes be handled? Would IL submitters even want encodes? Would it be too much work to encode every run that comes in, when ILs have the possibility to update several times a day in active communities?
[...]
IL Publication:
Encodes are not automatically provided to approved ILs, and are not required for approval as long as an input file is provided. Authors can opt in to having encodes made for them.
IL encodes made by us would go to a dedicated TASVideosChannel for ILs.
TASVideos may make backup encodes upon request, prioritizing author/community encodes as long as they are up.
Honestly, I think we should do a better job at emphasizing that TASVideos posting an encode on our own channel is an option and not a prerequisite for those submitting to our site. kierio has already mentioned a few communities that are reluctant to submit to us because users worry about having attention diverted from their own encodes elsewhere, and I think that even among our existing submitters there are some people who would have asked we link to their uploads if they knew that was allowed.
To talk about ILs specifically though, yes, I do like the suggestions being presented here. Whether or not the encode comes from the author (or whether it even gets an encode) should be up to them, and putting them on a secondary channel (and also following Riokaii's suggestion of waiting before encoding to prevent overloading staff with rapid improvements in a short timeframe) is probably a good way to go about things.
Samsara wrote:
[*]IL Moderation: Would managing ILs be too much extra work for us as staff? Could they be handled by community members? How do we determine eligibility for IL moderation?
[...]
IL Moderation:
Any staff member can "judge" and approve ILs.
There is no grace period for ILs, they can be approved as soon as they're submitted.
Community moderation is allowed for ILs: Each game can be assigned moderators, similar to SRC's moderation system, which would allow them to moderate ILs.
I think opening up the judging process of ILs to community leaders is a really good idea. Removing or dramatically shortening the 3-day grace period of full-game runs is also probably worthwhile, particularly given the more rapid rate at which ILs can be further optimised by other community members. I think this might also help address some of the uncertainty expressed by DigitalDuck as to how we ought to "define" ILs. I think trying to come up with an all-encompassing definition of ILs that specifies how finely they can or should be split up is inevitably going to run into problems with some game or another where it makes more sense to do something else, and so leaving these kinds of decisions up to community members on a per-game basis seems like the best solution, particularly where those communities are already well-established.
That being said, this does inevitably open the door to possible abuse of the system, where community moderators may enforce stricter guidelines about what kinds of categories, game versions, characters etc. are acceptable compared to what TASVideos does, and so I think it will be necessary to make sure things are being handled appropriately by those community moderators, or at least to present a clear line of contact through which users can report abuse to our core moderation team.
Samsara wrote:
[*]IL User Experience: How should ILs be displayed? How do we showcase them fairly in comparison to fullgame runs? How do we account for games that can only have ILs? How do we account for games that can't have ILs at all? Should every game have IL support by default even if there might not be distinct levels? Should ILs get submission and publication pages like fullgame runs? If not, how do we account for submission notes?
[...]
IL User Experience:
Any game with some way of identifying levels would get an IL table.
I think that part of the challenge with showcasing ILs fairly in comparison to full-game runs is that the way we present TASes is very different to SRC. While every game does have its own page from which publications and submissions can be viewed, what we really put more emphasis on currently is the per-platform publication lists, which is obviously not very helpful to ILs. Hosting ILs on game pages does make the most sense, but what we really need to do to avoid the challenges of "how do we account for exposure issues?" is give Games more priority on the navbar to better emphasise their importance.
Naturally there are going to be games where it doesn't make sense to have ILs - some games are only one level or one screen, after all - and it's difficult to know whether we ought to make the functionality opt-in or opt-out. Either way, we're going to have to manually correct hundreds if not thousands of games. Opt-in is perhaps better for the "community decides how ILs should be defined for each game" idea, though it will make getting word out about the new feature and how to use it more challenging.
I like the idea of using tables and sub-tables to showcase ILs and their different categories. I think that users should also be able to click on an IL to view the publication history for that specific level, and that this should be the primary way of showcasing work that has since been improved. I think that we should have submission pages (particularly for notes), but maybe not publication pages - writing up a description and screenshot for every level is asking a lot. Just linking to the necessary encodes and downloads should be sufficient.
Samsara wrote:
[*]IL Verification: Given the nature of ILs, can we justify changing the verification standards for them assuming they haven't already been changed universally before IL implementation? That is, could we accept them using savestates or provided save files without needing verification movies? Just for archival/historical purposes, could we go as far as not even requiring input files, assuming we'd still heavily prioritize the ones that have them?
[...]
IL Verification:
ILs can be approved with only a video, but will be marked as Unverified on the list. Unverified runs are hidden by default.
ILs are not required to have verification movies.
ILs are allowed to start from savestates.
While I think requiring input files for new submissions makes sense, we are introducing a feature to our site that already has an incredibly extensive history. I think it makes sense to try and represent that history on our site in some way, particularly the per-IL publication histories (we don't want to send the message that the decades of IL work done off-site is meaningless). And that might not always have a publicly available input file! Even for users who are still active now, it's inevitable they might have lost something in the years since from changing PCs, online storage, discords, etc. Having some kind of tag or symbol that says "unverified" or something might be a good way to compromise for these older runs.
ikuyo wrote:
There's a much larger conversation to be had about SaveRAM since it interacts with full game runs as well, and I want to reach a solution that reaches a good compromise between allowing people to make movies in the context they would be most comfortable with and making sure runtime environment is as reliable as possible. It is, in my opinion, the single largest issue to solve for our current methodologies, and I wonder if any attempt to make it work will allow our system to survive.
Agreed. Our system of requiring verification movies for absolutely everything is already a bit of a strain for games and categories that require tens of hours of play (or even more) to set everything up optimally, and the introduction of ILs is only going to put further pressure on the current ruleset forcing everyone to jump through these hoops. A more nuanced way of denoting what checks a particular IL or movie has passed along the lines of what Memory suggested is probably a compromise worth trying to begin with, with further adjustments being made overtime as deemed appropriate. Overall though, I think we do have to start accepting provided save files in these situations, and possibly also savestates (though those inevitably leave more room for tampering).
While we are already working on implementing alternative timing methods, I think it's also worth emphasizing that those are pretty crucial to us sufficiently serving the needs of the community with regard to ILs. How timing gets defined for ILs, even between different kinds of ILs in the same game, could also be left up to community members.
Before I forget, I would like to suggest allowing fastest lap tases for racing games where people are interested in those. I definitely know of some communities into that.
[16:36:31] <Mothrayas> I have to say this argument about robot drug usage is a lot more fun than whatever else we have been doing in the past two+ hours
[16:08:10] <BenLubar> a TAS is just the limit of a segmented speedrun as the segment length approaches zero
A reviewer (or some other role) describes its proposal. Judge reviews it and basically says "Yes" or "No". Also, some timeouts might be introduced. Like, if a submission is proposed to be rejected, it's not necessary to wait for a judge to reach it in a pile of other submissions.
(I can share this draw.io diagram if needed)
4. IL User Experience:
