OK, let's try this. Let's say next year's April Fools, masterjun submits a "96 exit" TAS that uses arbitrary code execution to do all 96 exits by teleporting to the end. Is this a correct obsoletion of the existing 96 exit TAS? Why or why not?
My post above should explain what the problem is (hint: the problem is that it shouldn't have been called '11 exits' in a category-labels-are-comprehensive world)
Well, the point is that the existing run must have a non-arbitrary reason to exist as a separate branch
And it does. It is the 'not glitched' counterpart to the 'glitched' run. Just that currently it is not well articulated how/why they are different/allowed to be different.
Either way, this version of "11 exits" with heavy glitch abuse is faster than the published "11 exits" with heavy glitch abuse, so go-go gadget publication.
Yeah, this.
It's hardly the first case that a newer run is faster but shows less of the gameplay than an existing run. That generally results in obsoletion of the slower run; I fail to see why this is suddenly a problem here.
My post above should explain what the problem is (hint: the problem is that it shouldn't have been called '11 exits' in a category-labels-are-comprehensive world)
So my bullshit detector is on the fritz, so I'm not sure how much of this thread is serious.
Either way, this version of "11 exits" with heavy glitch abuse is faster than the published "11 exits" with heavy glitch abuse, so go-go gadget publication.
The existence of this TAS as an april fools joke is to do with a bit of tasvideos history. Let me give you the summary:
How categories used to be named on TASvideos:
Arbitrary code execution, glitches to credits: 'glitched'
Forgoes arbitrary code execution, uses orb glitch and everything else, 11 exits: <nothing> or 'any%'
Then, a new ruling: "Let's not call categories 'glitched'. If it's the fastest category, we call it any%/<nothing>. Anything slower than that has to explain what goal or restriction makes it slower."
How the categories are named afterwards (I may have the steps out of order but it illustrates the problem):
Arbitrary code execution, glitches to credits: <nothing> or 'any%'
Forgoes arbitrary code execution, uses orb glitch and everything else, 11 exits: something like 'any% no arbitrary code execution'
Then it was pointed out by an SMW TASer: "Wait! What about the previous glitched TAS, that spat out a null sprite on a brown platform in Yoshi's Island 3 ('brown platform glitch' and jumped to credits early that way? If the categories are meant to be comprehensive about what the TAS must and must not do, then I could take the old obsoleted glitched TAS, submit it again and it would override 'any% no arbitrary code execution' even though it really shouldn't."
So it got renamed something like 'any% no stun sprite glitch, no brown platform glitch'
Then it was pointed out: "This is a really awful, verbose category name. It doesn't roll off the tongue, and people not familiar with the game's intricacies in its glitches won't have a clue what this category is actually about. More to the point - If you find another distinct game breaking glitch, then you have to update every other category and make the definition even longer. Eventually it won't even fit in one line."
So it got renamed something like 'any% no memory corruption'
Then it was pointed out: "Isn't the chuck eat/orb glitch memory corruption? It puts something invalid into a memory address via an invalid process."
So it got renamed something like '11 exits'
But we all know the problem with that - you're looking at the thread dedicated to it!
I suspect next it will be renamed something like 'any% no heavy memory corruption' (How do you define heavy vs light? An intuitive definition can work for a while, but only until a memory corruption glitch that isn't obviously heavy or light comes along.) or we will just give up and relinquish ourself to arbitrariness once more. (Also, what do we do with things like SMW 96 exit/SMW2:YI 100%, where lots of techniques that would make the run faster but more tedious/boring like L+Ring and chuck eating are banned? In an objective/comprehensive categories world it should become part of the category.)
(And if I'm missing any key points in this summary lemme know)
http://tasvideos.org/forum/viewtopic.php?p=323689#323689
Spider-Waffle demands that all doubters, libelers, and close minded non-productive people to the community will cease ALL activity related to OoT speedruns and TASes for 10 years.
Swordless is faster :D
This is already shaping up to be my TAS of the year 2014. I couldn't stop laughing at the OoB glitches, Superman looks so derpy doing them XD
Keep it up please
Are you pacifist as an extra goal, or because it never saved time to kill enemies in the route?
Movement questions:
3:00 Would you get to the shop faster if instead of bonking up against the corner you pulled right and just barely cut past the corner with as much leftward velocity as you could build up? (I mean, you can't fall faster or slower, may as well maximize X speed :D)
3:52: Why not jump off of blue robot 1 immediately? Bonking against blue robot 2 saves time? Is it an enemy movement manipulation?
4:11: Did you test 'megaman jump' style movement here? (The megaman jump is when, approaching a ledge that you want to fall down and then immediately change horizontal directions, you jump early so that you cut the corner with maximum downwards velocity without bonking on it, then change directions such that you cut the lower corner with maximum horizontal velocity and the extra downwards velocity as well again without bonking. Megaman jump also works when you need to fall down as fast as possible but don't need to change horizontal directions, e.g. when there's a sharp fall with walls on both sides you can't keep going horizontally until you've fallen a long way vertically)
4:24-4:25 feels sloppy but I can't tell if it is or isn't.
At 7:39 is it not possible to change your velocity so that you cut past the corner with the highest velocity you can rather than bonk on it (= no rightward velocity)?
At 7:55 could you not make the jump onto the higher second platform sooner so you don't bonk and lose horizontal velocity?
At 8:33 is there a reason why you take the corner so wide rather than just barely clipping by it with maximum rightward velocity? Is it an enemy manipulation?
At 8:52 it seems like you should cut the corner with maximum rightward velocity again. Or is there no acceleration underwater and there was no way to get a running start?
Btw - I loved all the tight enemy dodges and funny boss strategies. If it was a human player playing like that I'd be scared shitless XD But TAS can do what it wants~
If bug reports in the thread aren't guaranteed to be read, why even have a thread? Shouldn't the thread be one post: "Please go to the issue tracker: <link>" and locked?
Just so you know, someone on nicovideo is reuploading your Levitate% TAS but it's not marked as using the Levitate cheat code anywhere in the title or description.
http://www.nicovideo.jp/watch/sm23251335?ref=search_key_video
Someone who knows Japanese (Is FractalFusion in the house?) should PM the uploader and correct this misunderstanding.
He got a faster time recently.
Link to video
I think there is a new glitch (technique?) that prevents a boo from moving around while sucking it. Very useful for no OoB runs.
Regarding the branch name, this run was specifically done on a version of the game that didn't have everything unlocked, like EarthBound Zero with the breadcrumbs glitch. I think it would make the most sense to publish this run with the branch "no save glitch".
But you can't even do the " " glitch " " on this game. You can't ban something you can't do.
Please all speeds/ angles be right next to each other!! haha..
what about bytes when ram searching? is there a difference when searching betwen 1, 2, and 4 bytes?
If it's a float (like x/y/z in a 3D game), it's probably 4 bytes but could be 2.
If it's an angle it's probably 2 or 4 bytes.
If it's a small number, it's probably 1 byte.
Most of the Pokemon runs don't use code execution. Only rearranging some variables in memory using the game's existing inventory code. It's bordering on a nitpick, but they don't construct a program in memory and execute it.
Hmm, I was going to say 'it's just as strong, you can put them under the same umbrella' but then I remembered that 'Catch 151 Pokemon' category in Pokemon Red/Blue allows Item Underflow, just not arbitrary code execution using it. So you're right, the distinction is important (though only for Pokemon Gen 1, currently)
This run reminds me of Pokemon Yellow Warpless, which you finish in 6 minutes by performing arbitrary code execution then foregoing warps and just walking to (past...) the elite four.
Link to video
Pretty much means that, in any game with arbitrary code execution or similarly strong glitch, all other categories have to be stated in terms of banning glitches (instead of/in addition to/not just goals).
I also think that a distinction between light and heavy memory corruption IS possible and intuitive.
Or what if the input file was X frames late on one of the emulators? It's the same input, except you started one console a frame late - horrors, a desync! - but wait, it still works!
The difficulty in SMB1 glitchless is in what is a glitch.
To many people, jumping through a pirahna plant without dying looks like a glitch or cheat.
Or what about jumping through a hammer stream without dying? That technically is a glitch, because the hammer hitboxes are bugged and don't always match what you see on screen.