Joined: 3/9/2004
Posts: 4588
Location: In his lab studying psychology to find new ways to torture TASers and forumers
feos, very nice job on your updates.
Regarding "game end glitch", I'm wondering if perhaps we should turn these massive glitches into a glitch tag with a sub tag, meaning we could have:
glitch - game end
glitch - x-ray
glitch - pipe
glitch - coin case
etc...
Warning: Opinions expressed by Nach or others in this post do not necessarily reflect the views, opinions, or position of Nach himself on the matter(s) being discussed therein.
I kept thinking it all over, and here's the simplest I can put it:
1. The run uses what the game suggests as different gameplay options (amount of characters, use of warps, different characters or endings). Neither is implied, since either can appear faster.
Call out when:
2. The run uses the self-imposed conditions not suggested by the game as options (pacifist, walkathon, arbitrary code execution, playaround, 100%). Not using is implied, since otherwise the labels will blow.
Call out when:
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Joined: 3/9/2004
Posts: 4588
Location: In his lab studying psychology to find new ways to torture TASers and forumers
feos, I very much like what you're saying.
Also, the "movie classes" on the site should perhaps in some way tie into the call outs.
Warning: Opinions expressed by Nach or others in this post do not necessarily reflect the views, opinions, or position of Nach himself on the matter(s) being discussed therein.
feos, very nice job on your updates.
Regarding "game end glitch", I'm wondering if perhaps we should turn these massive glitches into a glitch tag with a sub tag, meaning we could have:
glitch - game end
glitch - x-ray
glitch - pipe
glitch - coin case
etc...
I believe it must be like that:
If the run uses the game-breaking glitch (that breaks the game execution and also cuts down the length several times), it must be labeled as "X glitch". X is figured out like that:
- skips to ending: "game end glitch"
- warps throughout the game, or right to near the end: "warp glitch"
- corrupts save data: "SRAM glitch"
- can't be abstracted: invent a new name with the help of the forums. Sometimes it already exists.
Nach wrote:
feos, I very much like what you're saying.
Also, the "movie classes" on the site should perhaps in some way tie into the call outs.
Movie classes serve for litstings, so should do branches.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Sounds good, I agree. I suppose that "best ending" or "julius mode" would also fall under this.
2. The run uses the self-imposed conditions not suggested by the game as options (pacifist, walkathon, arbitrary code execution, playaround, 100%). Not using is implied, since otherwise the labels will blow.
I agree with most of that, but I don't think that arbitrary code execution is a self-imposed condition. Rather, the opposite (no ACE) is a speed/entertainment tradeoff, so that would be a self-imposed condition.
2. The run uses the self-imposed conditions not suggested by the game as options (pacifist, walkathon, arbitrary code execution, playaround, 100%). Not using is implied, since otherwise the labels will blow.
I agree with most of that, but I don't think that arbitrary code execution is a self-imposed condition. Rather, the opposite (no ACE) is a speed/entertainment tradeoff, so that would be a self-imposed condition.
If a "speed-entertainment trade off" is used in all branches but one, why you and some other people so eager to see "no arbitrary code execution" in freaking all the other branches? And why do you love to blow the labels?
SNES Super Mario World (USA) "warps, no arbitrary code execution" in 09:57.82 by bahamete, Kaizoman666, Mister & PangaeaPanga.
SNES Super Mario World (USA) "game end glitch" in 01:39.74 by Masterjun.
SNES Super Mario World (USA) "small only, no arbitrary code execution" in 1:18:23.22 by PangaeaPanga.
SNES Super Mario World (USA) "96 exits, no arbitrary code execution" in 1:14:37.63 by bahamete, Kaizoman666 & Masterjun.
SNES Super Mario World (USA) in 02:25.19 by Masterjun. (this one uses ACE. and it's not a speedrun. it doesn't even beat the game. but looks like any%)
Really? Like that????
On the matter: does the game suggest you to either pick "arbitrary code" or not, as a normal player? If it doesn't, it's your own, self-imposed condition - to either pick it or not. I'm not telling using it is self-imposed, or avoiding it is self-imposed. I say the option itself is so.
Then, it becomes ONLY a matter of clean branches: we wither call out the use of that technique (that's so exceptional it does need a label, to show the viewer what it does) or we call out that it's voided (if it's so common the labels will become cleaner without it).
feos wrote:
1. The run uses what the game suggests as different gameplay options (amount of characters, use of warps, different characters or endings). Neither is implied, since either can appear faster.
Radiant wrote:
Sounds good, I agree. I suppose that "best ending" or "julius mode" would also fall under this.
Uh...
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
SM with X-ray glitch would be "X-Ray" or "X-Ray glitch"? X-ray can be used without glitching.
According to rule 1, X-Ray is most certainly not descriptive enough as you pointed out. The latter may also not be if there's more than one kind of X-ray glitch. So perhaps it should be something along the lines of X-ray bound breaking
As long as X-Ray glitch is complicated in its results, and quite similar to Crash Bandicoot 2's box glitch in that, and neither skips to ending, their meaning can be put into movie description, and the label be "X-Ray glitch" and "box glitch", telling it's still a game-breaking glitch by its wording.
It strikes me that most people, even fans of the game, may not know what a "box glitch" really is. So it may be a good idea to instead name the effect instead, e.g. "out of bounds". That would be clearer.
Aside from that, we have an interesting situation now over at Battletoads. Specifically, the "warps, 2 players" run is obsoleted by "warpless, 2 players", whereas the "warps, 1 player" run is obsoleted by "warps, 2 players". If I understand the system correctly, there should be five distinct branches here, i.e. "warp 1p", "warp 2p", "warpless 1p", "warpless 2p" and EGG; so perhaps some chains should be switched here.
If a "speed-entertainment trade off" is used in all branches but one, why you and some other people so eager to see "no arbitrary code execution" in freaking all the other branches? And why do you love to blow the labels?
Calm down dude, no need to get personal.
As Nach wrote, "2) All significant differentiators should be tagged where applicable. Even if this means every run for a game now has several tags.". To me, that suggests that all of the Super Mario World runs should be tagged as "warps" or "no warps", for example, or that this run would be "princess only, warps".
It strikes me that most people, even fans of the game, may not know what a "box glitch" really is. So it may be a good idea to instead name the effect instead, e.g. "out of bounds". That would be clearer.
Oh man...
feos wrote:
- skips to ending: "game end glitch"
- warps throughout the game, or right to near the end: "warp glitch"
- corrupts save data: "SRAM glitch"
- can't be abstracted: invent a new name with the help of the forums. Sometimes it already exists.
If the glitch is so weird and unused elsewhere that it can't be abstracted, it would NEED a name. Explaining it must be done in the movie description (and is done so in all such cases that I didn't miss).
Radiant wrote:
Aside from that, we have an interesting situation now over at Battletoads. Specifically, the "warps, 2 players" run is obsoleted by "warpless, 2 players", whereas the "warps, 1 player" run is obsoleted by "warps, 2 players". If I understand the system correctly, there should be five distinct branches here, i.e. "warp 1p", "warp 2p", "warpless 1p", "warpless 2p" and EGG; so perhaps some chains should be switched here.
Back then (look at the dates) no one cared about that. I'm not going to "fix" the mistakes that weren't mistakes. I only know that mine and MESHUGGAH's Battletoads obsoleted 1p warped because it was supposed to be the fastest branch.
Radiant wrote:
As Nach wrote, "2) All significant differentiators should be tagged where applicable. Even if this means every run for a game now has several tags.". To me, that suggests that all of the Super Mario World runs should be tagged as "warps" or "no warps", for example, or that this run would be "princess only, warps".
It tells to label all runs that can be labeled. As in, basing on the in-game options used that have counterparts, or on player-invented conditions. If there is a counterpart for "princess only, warps" - "princess only, warpless", they will both have these labels. If Battletoads "game end glitch" existed for both 1 player and 2 players, they you be "GEG, 1 player" and "GEG, 2 players". Same about Super C.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
If the run uses the game-breaking glitch (that breaks the game execution and also cuts down the length several times), it must be labeled as "X glitch". X is figured out like that
People spent many posts arguing about this and the poll doesn't show a majority for this option, so I suggest we don't go there again and instead stick with Nach's summary of what was decided in IRC.
Tell me where the poll gives the "uses x glitch" option.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Joined: 5/2/2006
Posts: 1020
Location: Boulder, CO
Late to the party, but I cast my vote for any% and "no x glitch" because it is the only scheme suggested here that would prevent confusion with regards to if one movie obsoletes another.
"glitched" does not do this because the trend for runs like the "11 exit" run seems to be that the runs uses some game breaking glitches, but not others, and its definition just changes in such a way that newer, glitchier runs aren't allowed to compete.
So I realize that this is already largely decided, and that this post is mostly reiterating what has been posted already, but I just wanted to say my piece.
Tell me where the poll gives the "uses x glitch" option.
I don't see how that question is relevant.
Look, my point is that you're moving too fast and making too many changes at once. Let's discuss in this thread how the various runs of Super Mario World should be named (following Nach's scheme), and once there is general agreement on that, then there is a default naming pattern that can be spread to other games.
I was thinking the whole day, trying to make it more universal. Here:
Movie branches exist to tell the viewer what approach the player used while TASing the game. There can be 3 foundations an approach is build upon.
1) Something that the game directly suggests (from menu, for example).
2) Something the game just allows (may have some indication though).
3) Something the game shouldn't allow, but it does (an erroneous assumption, a bug).
The use of these might need a label, to know if it does, we must answer a question:
Is the applied approach so common that the opposite is an exception?
If it is that common, we don't label runs that do it the common way, and label runs that don't, if there are counterpart runs of the same game.
If it's not that common, we label each approach, if there are counterpart runs of the same game.
When assessing range, something must be considered possible unless we are sure it is not. Obsoleted movies should count.
That's it! Now let's try some application.
Should we label runs that use Backwards Long Jump?Let's find the range: It only can be used in Super Mario 64.
Check the common way: Most of the SM64 runs use it (including all the obsoleted ones), a single one avoids it.
Result: When it's avoided, it must be labeled, if there are counterpart runs.
Should we label pacifist runs?Let's find the range: All games where one can kill enemies.
Check the common way: Among 882 movies (Shooter and Platform genres) there are 5 pacifist runs. 0.5%. Pacifist is exceptional.
Result: When it's a pacifist, it must be labeled, if there are counterpart runs.
Should we label 2/1 players in a multiplayer game runs?Let's find the range: All games that allow more than 1 simultaneous player (dunno how to list such games).
Check the common way: 32 use 1 player in a multiplayer game, 84 use 2 players. 38%. Neither is an exception.
Result: When it's 2/1 players in a multiplayer game, it must be labeled, if there are counterpart runs.
Applying it to "arbitrary code execution" or "game end glitch" gives the result: "must be labeled when used, if there are counterpart runs".
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Joined: 3/9/2004
Posts: 4588
Location: In his lab studying psychology to find new ways to torture TASers and forumers
I gave some thought to the idea of flagging the fastest run (the world record) with an icon of sorts. In the end, I dislike it, or rather, I see a bunch of issues with it.
The issue is that the fastest run is really subjective.
For starters, let's take Super Mario Bros., you can have the fastest possible SMB run completing only 8 levels, or you can complete all 32 levels. In truth, both are the fastest runs for their objectives, and each has its own world record. To prove the point, we have a Vault Tier which takes fastest any% and fastest 100%.
This same point applies for singling the fastest run which uses a single player or multiple players to complete a game. Both can be equally deserving over their own recognized record, even if one branch is faster than the other.
The next problem can be demonstrated with [2380] SNES Super Mario World "game end glitch" by Masterjun in 01:39.74 vs. [1944] SNES Super Mario World "warps" by bahamete, kaizoman666, Mister & PangaeaPanga in 09:57.82. Which is the record? There's two ways one can look at this:
1) [2380] SNES Super Mario World "game end glitch" by Masterjun in 01:39.74 is the record holder because it's the faster of the two.
2) [1944] SNES Super Mario World "warps" by bahamete, kaizoman666, Mister & PangaeaPanga in 09:57.82 is the record holder because the above movie is disqualified because it uses arbitrary code.
Alternatively, based on how one defines the ending, executing the credits or defeating Bowser and rescuing the princess will also determine which should be considered the record holder.
Since records are in truth subjective, I don't think we should necessarily be adding icons to show some runs are records and contrasting that others aren't. In fact, in most cases, when you see just two runs with clear labels, it's clear what a record is and you can use your own common sense or feelings. If we do need some indicators when there's a bunch of branches, perhaps we should consider counter imagery for branches which definitely are not considered any kind of record, for example, playarounds. Although, personally, it seems rather obvious to me that with the right labeling, that playarounds and similar are not record holders either.
Warning: Opinions expressed by Nach or others in this post do not necessarily reflect the views, opinions, or position of Nach himself on the matter(s) being discussed therein.
I think that post by Nach tells why "fastest possible game completion" can be implied (by some people), but can't be effectively used as the blank branch.
adelikat and Nach agree with me on this: blank branch =/= any%. This is why there was an agreement on IRC to call the "11 exits" (formerly "") branch of Super Mario World "warps". And to call out the used glitch for Masterjun's skip to game end. The fact that they are or were fastest among other branches doesn't make that "default" and doesn't justify the blank branch. Because any of them can be legitimately "any%". Which won't be used anymore due to complete ambiguity. The term "glitched" won't be used either for the same reason: it's too subjective and variable.
Blank branch, however, will be used. Just for a little different purpose. It won't mean any% in all possible cases. It will only mean that the run avoids everything that all the other branches use (unless there is no other branches). In quite many cases it would still fit both, but if it doesn't we shouldn't force it to become any%=blank label.
Here are the examples of the system, it was agreed about on IRC, but it needs some feedback. Please don't repeat "blank branch is the only sane way to label the fastest branch", read the above Nach's post why it can't always be so.
http://tasvideos.org/Game/nes-super-mario-bros.htmlhttp://tasvideos.org/Game/snes-super-mario-world.htmlhttp://tasvideos.org/Game/n64-super-mario-64.htmlhttp://tasvideos.org/Game/snes-super-metroid.htmlhttp://tasvideos.org/Game/nes-battletoads.html
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
I want to say you can just the word "warps" because it is common sense.
We currently the word "warps". As for common sense, different people appear to have it different. But who is laborious enough will be able to improve the system over that to overcome the contradictions.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Joined: 10/12/2011
Posts: 6337
Location: The land down under.
feos wrote:
We currently the word "warps".
Herp. Derp. I didn't realize I removed what I had.
I want to say you can justdropthe word "warps" because it is common sense.
That explains why the second comment doesn't make sense.
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account.Something better for yourself and also others.
Exactly what I said. To me common sense is to use it.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
The issue is that the fastest run is really subjective.
tasvideos.org has become a kind of de-facto "authority" of all things tool-assisted via the fact that it's the biggest and most popular TAS website out there (which is a great thing.) Thus it's most likely the source that people will use to get information on tool-assistance and tool-assisted speedruns.
Let's assume that someone is eg. writing an article on tool-assisted speedrunning; this might be even some publication of great prestige (perhaps all the way up to such notorious publication as the Guinness Book of World Records, some day perhaps... we can dream). If they are looking for the "world record" completion time for a certain game (perhaps to compare it to the unassisted run, or simply to list it as such), it would be good to clearly mark the branch that's our official such record.
With the vast majority of games the record time is quite clear, in most cases because there is only one TAS of it (or at most two, the other often being clearly longer and labeled with something like "100%"). With other games it can be less clear, if there are numerous branches and their goals are less clear to the average viewer.
What will most likely happen is that the visitor will either choose one of the branches pretty much at random (because they might even be unaware that there are other branches, so they will choose the first one they stumble across), or simply the branch with the lowest time. In many cases the latter will probably be ok (after all, if it has been published it ought to conform to the publication rules, and if it has the fastest completion time, it is the "world record".) However, it may be that in some cases the fastest completion time uses some odd technique that, while not disqualifying it from publication, may "disqualify" from being the "official" word record. (Also, some games just don't have a "world record" completion time because of their nature. They might only have a playaround that doesn't even try to complete it as fast as possible using any means possible.)
(Of course if we marked world record runs somehow, it would also be necessary to make it clearer that such marks exist and what they mean.)
However, it may be that in some cases the fastest completion time uses some odd technique that, while not disqualifying it from publication, may "disqualify" from being the "official" word record.
Like which?
Also, about someone writing an article, if the person understands what he writes, he will not pick up the unassisted 100% run of Chrono Trigger to compare it to the mid-frame reset TAS, right? He wouldn't even compare the mid-frame reset TAS to the non-mid-frame reset RTA. Because his senses will tell him: "Man, the difference can't be several hours within the same gameplay conditions. RTA can't be that much slower within those conditions. I need to pick the comparable branches."
Ok, so he decided that. Now, while looking at RTAs, he will (in most cases easily) see what were the general conditions: warps, players, ending, difficulty, character. If he looks at TASes, there will now be labels for all of those (whenever it may confuse someone to not have them). So it will make the comparison the easiest possible.
And if there's only one branch, we will need to know what might disqualify the run in someone's eyes. Like, some people consider using Backwards Long Jump in SM64 and avoiding it similar conditions. Some may think it must not be used, and its use disqualifies the "world record" (say, because it can't be used in real-time, or whatever). Some think that avoiding it isn't a world record in any sense. This is why "world record" is subjective.
But in all subjective things there is still a way to get the statistics. If they show that avoiding BLJ is only done in 5% of all submitted SM64 TASes, it's should be considered an exception and labeled when avoided. If it's ~40% of all published TASes of SM64 - I'd label both cases.
BTW, note that in my latest iteration of the system, I never used anything pre-defined, like "game-breaking glitch". It's because such definitions are in fact very subjective. And no matter in which way these techniques are used, only statistics will tell how to label things with the most advantage.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
About Arbitrary Code Execution and all related stuff.
If we use ACE with subtags, there will be 2 situations.
- ACE used primarily, for the sake of itself, to make some glitchfest playaroiund.
- ACE used only as a mean for something other, like ending the game instantly, or warping within each level to its end, like Masterjun did.
For the first case, I suggest switching to the term "total control". Because it doesn't only take total control over the game and then just drop it as a sacrifice for speed, but it keeps applying it more and more, showcasing the very fact that total control is taken.
In the second case, I suggest not using any term at all (ACE, ECA, TC), only the primary goal it's used for - game end skip, or whatever else is considered Moons.
That way total control branch would be so rare that it wouldn't need any sub-labels yet. It only would, if there are several such runs per game. Then, to differentiate between them , we would actually need some sub-label. "total control: Pi", "total control: Ponies", "total control: new games" or whatever. As long as TA exists for different games only, we don't have the need to sub-label it.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Joined: 3/9/2004
Posts: 4588
Location: In his lab studying psychology to find new ways to torture TASers and forumers
feos wrote:
About Arbitrary Code Execution and all related stuff.
If we use ACE with subtags, there will be 2 situations.
- ACE used primarily, for the sake of itself, to make some glitchfest playaroiund.
For the first case, I suggest switching to the term "total control". Because it doesn't only take total control over the game and then just drop it as a sacrifice for speed, but it keeps applying it more and more, showcasing the very fact that total control is taken.
That way total control branch would be so rare that it wouldn't need any sub-labels yet. It only would, if there are several such runs per game. Then, to differentiate between them , we would actually need some sub-label. "total control: Pi", "total control: Ponies", "total control: new games" or whatever. As long as TA exists for different games only, we don't have the need to sub-label it.
It's hard to predict the future, but say Masterjun made the following:
Super Mario World now has the Tanooki suit from Super Mario Bros 3, you get it every time you hit a Yellow switch block. It also has the Hammer Bros. Suit which you get from Green switch blocks. And this in-game hack is now used for a playaround to go crazy within the game completing 96 levels in the most absurd manner possible.
The above would be quite different than what you want to currently label as playaround. What would you label this case?
You suggest we only more descriptively subtag label runs in case of more branches being created for a single game. I disagree, certain subtags will end up carrying connotations as to how they're used. Playaround right now is used across the site to suggest a long winded but insane abuse of the games psychics engines, bug demonstrations, incredible activities, and more. I would not consider adding on new games or a pony cutscene to be the same category.
We should aim for subtags which work across all games, and choose them up front.
Warning: Opinions expressed by Nach or others in this post do not necessarily reflect the views, opinions, or position of Nach himself on the matter(s) being discussed therein.
Uhm... What? Am I the only person that sees a problem with this?
TASing (and with that, TASVideos) has always been about 1 thing: creating an as short as possible input file that, when played back, reaches the end of the game.*
It still seems to me that with the advent of ACE (/TC) and suddenly glitching to the end of the game this entire idea has been abandoned. It's as if a bunch of people suddenly thought "Oh, we don't particularly like this anymore, better apply some arbitrary criteria and call it 'glitched' and then stop considering it as proper TASes!" This is ridiculous, these runs are the fastest, and objectively so. You might not like them, but I don't give a shit.
* I realize that this is not the sole goal of TASing, but it has always been the sole criteria for Tool-Assisted Speedruns.
