This is a TAS that completes all 96 exits in Super Mario World using arbitrary code execution. I wrote some code in the game's memory that allowed me to give Mario maximum speed, both left and right.
The process of how the code was written in memory is the same as my glitchfest with arbitrary code execution, so I won't go into details here.
The code was very simple: pressing L on controller 1 sets Mario's speed to 7F and pressing R sets the speed to 80. The address responsible for Mario's speed is $7B and the maximum values ​​that we can have are those mentioned (7F is the maximum to the right and 80 is the maximum to the left). In short, Mario travels to the right at a speed of 127 pixels per frame and to the left at 128 pixels per frame.
The code is this:
LDA $18
AND #$10 ;|check if R (controller 1) is pressed
BNE ace1
LDA $18
AND #$20 ;|check if L (controller 1) is pressed
BNE ace2
RTL
ace1:
LDA #$80 ;|set 80 in $7B if $18 equals 10
STA $7B
RTL
ace2:
LDA #$7F ;|set 7F in $7B if $18 equals 20
STA $7B
RTL
In short, that's it. No addresses other than the code have been changed.
Important things:
Mario jumps very high when his speed is between 110 and 119 (decimal). I used this to my advantage;
At high speeds, Mario can pass through walls without dying. You will see this several times in this TAS.
This may be a bit of a controversial category because of the use of arbitrary code execution, but I find it fun to watch Mario blast through every level like a lightning bolt! I also believe that there are several improvements in this TAS and it could be faster, but I didn't have much free time to think about the best possible strategies for each level.
My thanks to the great Noise de Gole for teaching me a lot about arbitrary code execution and you for watching :)
feos: At the end of the day I can't consider this movie a playaround. If it was, there would be some combination of changing the existing game with some entirely new payload, and both would have the goal of being subjectively exciting in their creativity. Creativity is the whole nature of the playaround goal.
The goal of this movie is on the contrary rather objective - getting all exits as fast as possible - but the game itself is tweaked to help with it... in a subjective way.
Playarounds are judged by entertainment, and speed oriented goals need to be clear and solid to go to Alternative. But the part where ACE is used to hack the game to be faster is not very solid, according to my view and to the viewer feedback that this submission got.
There's no definitive way to hack this game to make it optimal for the "all exits" speed goal: as long as we have total control, we can force the game to only execute code that we consider as "an event of using an exit", but there's no way to clearly define that even on the game code level! If it's just about internal flags, then setting them directly via ACE is not allowed for "full completion", but this movie is not "full completion" either! So we arguably should allow setting those flags, and then there's nothing else that would have to happen in the run. Or if we count "passing through an exit physically" as "an event of using an exit", then we also remember that inside the game every such event is compound, and it's only up to us how exactly we want to slice it - in order to only execute that part.
So since this doesn't work as a playaround or as a speed goal, the only class I can send this to is Playground.
I want all good TAS inside TASvideos, it's my motto.
TAS i'm interested:
Megaman series, specially the RPGs! Where is the mmbn1 all chips TAS we deserve? Where is the Command Mission TAS?
i'm slowly moving away from TASing fighting games for speed, maybe it's time to start finding some entertainment value in TASing.
Because there are only 256 possible values ​​for each address in the game. In the case of the address responsible for Mario's speed, it is divided into 2 parts (because Mario goes left and right), so we only have 128 possible values ​​for each side. Therefore, 128 is the maximum speed, both left and right.
I'm not sure but should this obsolete the regular 96-exits run? Idk but it feels like it should cause they're close in time, but again this is an ACE version of the category, so a lot is different content wise
Nonetheless, good run! Yes vote!
It is fun to watch, not a bad run, but...
While you're at it, why not make the screen transitions faster also?
Why not make overworld travel faster? Please max the speed of overworld travel!
This is what I would like to see: Max the speed of everything. This includes enemies and everything else that moves.
Not a bad run, but not THE FINAL FORM!
I believe it can be made much more amazing than this...
Please think of ways to make the concept more interesting. XD
Pre-emptively voting yes for an ACE run that maxes the speed of EVERYTHING.
It is fun to watch, not a bad run, but...
While you're at it, why not make the screen transitions faster also?
Why not make overworld travel faster? Please max the speed of overworld travel!
This is what I would like to see: Max the speed of everything. This includes enemies and everything else that moves.
Not a bad run, but not THE FINAL FORM!
I believe it can be made much more amazing than this...
Please think of ways to make the concept more interesting. XD
Pre-emptively voting yes for an ACE run that maxes the speed of EVERYTHING.
I had a similar thought. If you're already maxing out Mario's horizontal speed, why not also max out his vertical jumping acceleration/speed? Wouldn't that make vertical stages faster, and potentailly minimize/negate the need for the cape as well?
EDIT: Still a fun watch, as is.
It is fun to watch, not a bad run, but...
While you're at it, why not make the screen transitions faster also?
Why not make overworld travel faster? Please max the speed of overworld travel!
This is what I would like to see: Max the speed of everything. This includes enemies and everything else that moves.
Not a bad run, but not THE FINAL FORM!
I believe it can be made much more amazing than this...
Please think of ways to make the concept more interesting. XD
Pre-emptively voting yes for an ACE run that maxes the speed of EVERYTHING.
Man, that's a great idea, but I don't know if I would have enough knowledge to do something like that. I'll study the possibilities and maybe make this TAS obsolete soon hehe
I'm not sure but should this obsolete the regular 96-exits run? Idk but it feels like it should cause they're close in time, but again this is an ACE version of the category, so a lot is different content wise
Nonetheless, good run! Yes vote!
No, I think they are very different categories. If this TAS is accepted, it would have to be in a new category. It is not fair to compare a "normal" 96 exit with one made with ACE.
Joined: 8/15/2005
Posts: 1946
Location: Mullsjö, Sweden
If the goal is to beat all stages as quickly as possible while using ACE, then I would say that this run is greatly suboptimal, considering all of the things you could do. Limiting to just changing the speed in levels would in generally be a quite arbitary ruleset for this goal. Will thus vote no based on these two factors.
It sort of reminds me of the 100% Majora's Mask TAS using ACE, where the general 100% say that you need to get each item from its general source. Once the ACE setup starts, it's just warping around to get all the required items. As seen here: https://youtu.be/o7wGpmP6YH8?t=1480
I suppose something similar would also be possible here. As soon as one level is beaten, you instantly warp to the start of the next one, if the goal definition is to "Start and finish all stages". Then again, the definition what start and finish is would be debateable as well, for an ACE definition.
If the goal is to beat all stages as quickly as possible while using ACE, then I would say that this run is greatly suboptimal, considering all of the things you could do. Limiting to just changing the speed in levels would in generally be a quite arbitary ruleset for this goal. Will thus vote no based on these two factors.
It sort of reminds me of the 100% Majora's Mask TAS using ACE, where the general 100% say that you need to get each item from its general source. Once the ACE setup starts, it's just warping around to get all the required items. As seen here: https://youtu.be/o7wGpmP6YH8?t=1480
I suppose something similar would also be possible here. As soon as one level is beaten, you instantly warp to the start of the next one, if the goal definition is to "Start and finish all stages". Then again, the definition what start and finish is would be debateable as well, for an ACE definition.
Well, the goal I set for this TAS was to be as fast as possible using only the super speed code. I limited myself to this code only, because ACE has no limits.
I could simply, in the code itself, set all 96 exits and in less than 1 minute I would have the complete game, as I have already done here: https://youtu.be/SqoXvVKS9bk or put a code in memory that would allow me to complete the levels in the first possible frame, as I have also done here: https://youtu.be/EzoXb465Sto
This TAS is much more about entertainment (and maybe it would even be cool to change the branch) than about completing 100% of the game as quickly as possible using ACE.
But yeah, you are not wrong in your point. In fact, there are hundreds of ways to be much faster than the way I chose to do it.
I think the run's goal is not as arbitrary as it may seem. Sure, with ACE everything can be done. The point is to find interesting things to do with it.
Maximizing the movement speed of the player character is less arbitrary than maximising the movement speed of some other object in the game.
Some games like Doom have setting where you can set the maximum player speed to 255. ACE can be used to emulate such setting. That is pretty much what this run does.
What I would like to see is a Doom TAS where enemy speed is also set to 255. Such setting does not exist.
Yes, to some degree it will always be arbitrary no matter what. But I would love to see a run where the speed for the player and enemies and all moving objects are maximised. For every game. It is a new category of TAS entirely. Arguably it is even more TAS than TAS.
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.
IgorOliveira666 said:
or put a code in memory that would allow me to complete the levels in the first possible frame, as I have also done here: https://youtu.be/EzoXb465Sto
This method should work too
I abstain from voting
Currently making a route for Super Bomberman 5 200%
TASes i'm planning: aero fighters, tg3000 (again), gradius 3, bust a move, bust a move plus, gradius rebirth, smg, smg2, mp9, gh3 (wii)
I don't have any plans to make other categories with ACE, but it wouldn't be a bad idea to create a hack. It would even make everything faster, including the map, enemies, bosses and everything else, as mentioned above.
And regarding the route, the biggest differences were in relation to Yoshi, because in my TAS, since I had super speed at all times, I didn't need Yoshi in several levels, like in Forest Of Illusion 2, for example. But I confess that I missed some wings and maybe even some stun glitches. Even with super speed, it would probably be faster, but I only stopped to think about it after I had already finished it. This TAS definitely still has a lot of improvements.
Joined: 4/17/2010
Posts: 11656
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg
Thanks.
So here's what I'm thinking about this movie. Since ACE gives the player absolute freedom, speed challenges become moot with that superpower. Normally for any kind of playaround the game is used as a canvas on which the author can draw anything that they're inspired to create, using in-game assets as brush and paint. But with ACE, there's no game anymore. In-game assets are arguably still there but they are meaningless because they can be arbitrarily changed or ignored. We're basically ROM-hacking live.
And with absolute freedom, the only challenge that makes sense IMO is to be as impressive as possible, however broadly we define being impressive. Classic ACEplayarounds use in-game assets to create a spectacle, and the result was very entertaining as well as technically stunning.
Then there's a submission that streams video from controller inputs, but it's an ACE playaround for a game that already has an ACE playaround that's impossible to one-up. Playarounds are by nature judged by entertainment, and it'd be impossible to obsolete the existing one, but there's some talk that ACE is such a cool thing that any game can have a "bad apple" movie published, it just never gets old.
So feedback here was mostly positive but people were also agreeing that with ACE any speed limitation is arbitrary. Intuitively, the only way to approach this submission is still treating it like a playaround, and any other potential speed goal using ACE may have a chance to obsolete this. But then we look at #4252: Masterjun's SNES Super Mario World "11 exits" in 08:07.53 and it's kinda the same concept there but taken to an extreme, except that run uses an arbitrary number of exists instead of all of them. But it reaches them much faster!
And there's also [6035] SNES Super Mario World "arbitrary code execution, glitchfest" by IgorOliveira666 in 42:49.44 where I said:
feos wrote:
What I'm unsure about is whether it makes sense to have several movies with the same payload type published separately, or is it better to split them only by payload type, or in some other way? Masterjun's ACE is an entirely-new type, and this submission is a hack of what is there. Or maybe we approach it from a different angle and just rename this movie to "ACE, glitchfest"?
This movie is "a hack of what is there", and on the surface it looks like it's neither a playaround nor a glitchfest. But as I said speed challenges are moot with ACE, so it kinda has to be either of those 2.
The only other point of view on this seems to be allowing all such variants freely, but they can't all live in the Alternative class, because there's infinite potential for different ACE sub-goals. There's one place for almost unlimited creative freedom tho - Wiki: Playground.
Thoughts?
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 definitely agree with what feos said here. In general, I think ACE is for entertainment/playarounds, or simply if you aim for speed, you just have to finish the asap. Using ACE on another speed category that already restricts itself from finishing the game too fast just feels like two contradictions to me. Like I already said, this and regular 96-exits run are close in time thus you're going fast, yet you still also want to get all exits on the way. This doesn't mean that this run is not well made or not appealing to watch, but it just feels like ACE isn't suited well for these categories. ACE allows you to control number of exits, right? Well, that is another thing that 96-exits wouldn't allow either. I voted yes for this run still but I guess it may be better suited for Playground.
Thanks.
So here's what I'm thinking about this movie. Since ACE gives the player absolute freedom, speed challenges become moot with that superpower. Normally for any kind of playaround the game is used as a canvas on which the author can draw anything that they're inspired to create, using in-game assets as brush and paint. But with ACE, there's no game anymore. In-game assets are arguably still there but they are meaningless because they can be arbitrarily changed or ignored. We're basically ROM-hacking live.
And with absolute freedom, the only challenge that makes sense IMO is to be as impressive as possible, however broadly we define being impressive. Classic ACEplayarounds use in-game assets to create a spectacle, and the result was very entertaining as well as technically stunning.
Then there's a submission that streams video from controller inputs, but it's an ACE playaround for a game that already has an ACE playaround that's impossible to one-up. Playarounds are by nature judged by entertainment, and it'd be impossible to obsolete the existing one, but there's some talk that ACE is such a cool thing that any game can have a "bad apple" movie published, it just never gets old.
So feedback here was mostly positive but people were also agreeing that with ACE any speed limitation is arbitrary. Intuitively, the only way to approach this submission is still treating it like a playaround, and any other potential speed goal using ACE may have a chance to obsolete this. But then we look at #4252: Masterjun's SNES Super Mario World "11 exits" in 08:07.53 and it's kinda the same concept there but taken to an extreme, except that run uses an arbitrary number of exists instead of all of them. But it reaches them much faster!
And there's also [6035] SNES Super Mario World "arbitrary code execution, glitchfest" by IgorOliveira666 in 42:49.44 where I said:
feos wrote:
What I'm unsure about is whether it makes sense to have several movies with the same payload type published separately, or is it better to split them only by payload type, or in some other way? Masterjun's ACE is an entirely-new type, and this submission is a hack of what is there. Or maybe we approach it from a different angle and just rename this movie to "ACE, glitchfest"?
This movie is "a hack of what is there", and on the surface it looks like it's neither a playaround nor a glitchfest. But as I said speed challenges are moot with ACE, so it kinda has to be either of those 2.
The only other point of view on this seems to be allowing all such variants freely, but they can't all live in the Alternative class, because there's infinite potential for different ACE sub-goals. There's one place for almost unlimited creative freedom tho - Wiki: Playground.
Thoughts?
I'm really confused. In fact, I agree with everything you said. Before posting this TAS here, I thought about putting it as "96 exit, ace playaround" or something like that, but I don't think that's very appropriate. Maybe creating a hack with this ACE code already inserted, as you said yourself, would be the best option. Regarding this one, do as you see fit.
Joined: 4/17/2010
Posts: 11656
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg
IgorOliveira666 wrote:
Maybe creating a hack with this ACE code already inserted, as you said yourself, would be the best option. Regarding this one, do as you see fit.
That would open room for publishing all the same goals as SMW has but in this fast environment. In the hacks I linked Acmlm changed a few more things on top of Mario speed:
Superfast Mario Bros. description wrote:
The framerule was changed from 21 frames to 1 frame, instant acceleration to 0x7F speed, various delays cut out or speed values increased, nearly instant level clear.
It’s about as fast as acmlm could make it go, and was meant mostly for a silly TAS, but turns out to be actually beatable too!
Such a hack would be a much more straightforward approach to such speed goals, without aiming at policy borderlines regarding ACE 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.
Maybe creating a hack with this ACE code already inserted, as you said yourself, would be the best option. Regarding this one, do as you see fit.
That would open room for publishing all the same goals as SMW has but in this fast environment. In the hacks I linked Acmlm changed a few more things on top of Mario speed:
Superfast Mario Bros. description wrote:
The framerule was changed from 21 frames to 1 frame, instant acceleration to 0x7F speed, various delays cut out or speed values increased, nearly instant level clear.
It’s about as fast as acmlm could make it go, and was meant mostly for a silly TAS, but turns out to be actually beatable too!
Such a hack would be a much more straightforward approach to such speed goals, without aiming at policy borderlines regarding ACE branches.
I don't think I have enough knowledge to make the entire game faster, but I can certainly make the map, Mario's speed, enemies, and auto scrolls much faster than normal, which would already be really cool.
