Hmm...I think I would have preferred watching the Marathon mode rather than the Time Trial mode, simply because the gameplay is more continuous. Any thoughts?
I haven't either; it's actually the title screen demo level from the original Demo World (that only contained 1 world). Apparently it was left in DWTLC in whatever level slot was supposed to correspond to the unique World 4 Dragon Coin Heaven (in the original SMW, that one Coin Heaven which was at nighttime and contained enemies). You're never supposed to reach it; the game doesn't have any normal Yoshi Wings. Note the glitched color palette in the HUD.
If it weren't such an obscure reference, I'd nominate using it as the publication screenshot, since those familiar with DWTLC wouldn't ever recall seeing that level, and those unfamiliar would still be scratching their heads wondering how Mario is apparently standing inside a wall.
"Does not eat any chucks"
What about Does not spawn any Reznor?
The same logic applies...
The logic behind the decision is that chuck eat -> bring orb to level -> use orb to skip level works regardless of what the level is. Whereas spawning a reznor to beat the level is interesting because it relies upon properties of the level you enter.
If you can spawn a reznor/skip the level almost anywhere in Super Demo World 120 exit, it would of course be banned under a similar mindset. It's kind of like how you're not allowed to use ACE in 100% style categories.
I'd say that Reznor spawning should be allowed since Yoshi Wings spawning is, and both are "partial skips" that can only apply to a portion of the levels.
---
On a different note, how exactly does the Cloud-Bowser glitch work, and how does it skip phases?
^Thanks! Could it reasonably be said that a run banning ACE would necessarily also have to ban execution reaching Open Bus to avoid arbitrary goal statements?
Also loving the fact that a TAS of a SMW-based game on this site is finally able to show off the Cloud-Bowser glitch without any ACE-related arguments.
The cloud strategy was originally discovered in SDW and had been meant to be used there. So this is rather a reverse import or something. AND I think this is a good reason for not using this trick in SMW; we have a more appropriate hack to use it, SDW!
Huh, didn't know about that first part.
And yes to the second part, although the primary reason is that from a superhuman perspective, in vanilla SMW, there's no non-arbitrary reason to get the cloud (forcing ACE to be possible since the game jumps to Open Bus) and then not use that ACE potential to instantly beat the game. The fact that the 'regular' cloud glitch is still able to be shown off is a really nice side effect bonus.
A Yoshi wings exit was found in Misty Star World by bahamete after I finished this run. This alone will save 3-4 minutes.
By "after I finished this run", do you mean literally between today and July 3 (so sometime in just the past 3 days)?
Also loving the fact that a TAS of a SMW-based game on this site is finally able to show off the Cloud-Bowser glitch without any ACE-related arguments.
Apologies if this is duplicating an earlier question, but:
When using the Chuck-Eat glitch to get a cloud, does the game always jump to Open Bus allowing for ACE? If so, are real-time speedrunners simply executing NOPs while the game returns to 'normal' mode? And would blindly spamming buttons during the "obtain cloud" time period most likely cause a bizarre crash?
Using the then-glitched inventory (which is now nothing more than a memory editor quite literally) to magically transport yourself from one map to another is no more complicated than using it to write programs to do things that weren't intended or accounted for (catching Pokemon at levels that can't happen/catching Pokemon that can't be wild period, etc).
Arbitrary code execution:
- an attacker's ability to execute any commands of the attacker's choice on a target machine or in a target process
Considering item underflow literally lets you do anything you want with the game, I think it's appropriate to consider it ACE.
There's no "appropriate" or "consider" in the definition. ACE stands for "arbitrary code execution", not "arbitrary memory editing". Item underflow allows the latter, but not the former except with much more elaborate setups (and if that setup isn't performed, no ACE for you).
Basically, it boils down to this one question:
Has the PC jumped somewhere it wasn't supposed to?
No? It isn't ACE. Simple as that.
Let me attempt to make this clear: Arbitrary code execution is defined as causing the game's program counter to jump somewhere it is not supposed to go. If the PC doesn't do this, ACE has not occurred. It doesn't matter how ****ed up the rest of the game's memory becomes; it's not ACE unless the game actually runs an instruction not programmed into it.
As far as I know, several 'extreme' glitches such as item underflow, save corruption, and ZZAZZ do heavily corrupt the game's memory, but they do not ever cause the game to run an instruction that was never programmed into it. Therefore, they are not ACE. Even several easy and well-known glitches such as encountering Missingno. and performing Trainer/Fly corrupt the memory, changing values in the game that weren't meant to be changed that way. But they do not perform ACE.
Using the "Cooltrainer" move in battle is ACE since it immediately causes the game to run instructions starting from the stored data of the last opponent's party. The PC was never supposed to go there. Similarly, in English Gold/Silver, opening the Coin Case at all (ever) is ACE, since it immediately causes the game to run instructions from a sound bank.
As for the current Cooltrainer glitch used in the real time "catch 'em all" speedrun, I do not know whether it is ACE or not. It doesn't ever select the move; it only displays the move's name on screen a bunch of times. I don't know whether the PC ever jumps to an invalid location. But if it doesn't, then it's not ACE.
There's already several Tetris TASes on this site; why not a TAS from arguably the hardest Tetris series of all time?
A counter-argument that I've heard is that realtime play is already at such a high level that a TAS would look pretty similar. I'd have to disagree; if you take a look at this example TAS from TGM2+:
https://www.youtube.com/watch?v=OZaq0jQcZVQ
even with the super-fast regular gameplay of Death mode, this run is able to cut the unassisted play time record nearly in half (~169 seconds compared to ~293). For more experienced watchers, a TAS could also perform tricks that are known to be borderline impossible at regular speed; the most straightforward example would be stacking with the Tetris hole at the left instead of the right, something which virtually zero realtime players attempt since it is extremely risky and offers no benefit.
Any thoughts?
I'd personally prefer a multi-run to actually have the inputs visibly synchronized; having too much of "left+right or down+right can do different things on different games" as in the current publication almost makes things unremarkable. This could be further improved by making the input display visible somewhere in the encode.
It might also be entertaining to have SMB1's second quest finally completed for once, as no current publication does that.
^So it's like a hardware glitch based on undefined behavior?
If so, I'm echoing an earlier sentiment in that I'll wait for console verification to vote.
Why is it necessary to use a coin (or specifically, have Mario collect the coin and have Yoshi catch it with his tongue on the same frame)? Is it part of the Chuck-eat glitch setup?
Does jumping into Open Bus depend specifically on a running chuck? I.e. I'm guessing a jumping chuck gets a goal orb while a running chuck gets a Lakitu Cloud; is that right?
Basically even though I understood most of the previous SMW runs I'm not sure what's going on here.
^But that would defeat the entire purpose of the category if it should exist, which is explicitly to find the minimum number of A presses necessary to collect all the stars. It would be like the "small only" Super Mario World TAS using something other than small Mario for an exit, even if the strategy for that exit somehow took hours to set up as small Mario.
This sounds like a situation from earlier with the SMB1 lowest score TAS. Sure, it's very entertaining to see how to avoid all those points, and it's fine when you can cut out the 60-second+ wait at the flagpole. But when you have make it into a full uncut movie, watching Mario wait and dance around over and over repeatedly can get boring.
Am I saying that a minimal A-button TAS would inevitably be rejected on this site if submitted? No, of course not. But the fact that improvements are constantly being found much faster than in other TASes, and that some of them (more than this one) result in the movie's length being extended by several hours, make it not too likely to be published or even submitted in the first place.
The item duplication on its own isn't terribly interesting given that it's already possible with Pokemon cloning but has the negative side effects of filling your boxes with Bad Eggs and possibly trapping you, but here's a critical thing: it seems like the farther you scroll, the later the corrupted addresses are (assuming that items are stored after Pokemon data as in Gen I). This theoretically means that if you scrolled for a while just right, would it be possible to switch your current location to the Hall of Fame and cause a glitch warp as in the RBY movies?
Neither of those force the PC (program counter) to jump to a location like the save data/party data/controller input that can be freely handled by the player; item underflow in particular doesn't move the PC at all. So they aren't arbitrary code execution.
Was taking a look at some real time Pokemon speedruns recently and noticed this:
http://wiki.pokemonspeedruns.com/index.php/Pok%C3%A9mon_Red/Blue/151_Pok%C3%A9mon
The current real-time record is about 3:29, only 9 minutes short of the current TAS record. This was done using the help of several glitches which were undiscovered at the time, such as item underflow and the CoolTrainer move. This makes me believe that the current TAS record could be massively improved using these new glitches.
Admittedly the boundary between what glitches are allowed and aren't allowed is a bit arbitrarily chosen, although it can be fairly well-defined ("Pokedex 'caught' flags must be set by the game and not corrupted; also no arbitrary code execution"). Given that this site's "completion" requirements are generally defined by the speedrun community (e.g. the Silver run must beat Red and not just the Pokemon League), and arguing that said Silver run's goals are slightly arbitrary anyway (in that you have to see Red in Mt. Silver and talk to him rather than just run the credits routine), would anyone seriously consider attempting an improvement to the "Gotta Catch 'Em All" TAS?
Writing to any even location in the range 8000-9FFE will program the MMC3 Bank Select register. Bit 6 of this register selects the PRG bank mode. When set, the first 8K is fixed and the third 8K is swappable. When cleared, it is the other way around. SMB3 uses the layout where the first 8K is fixed. By writing 80 to 9C70, the other layout is selected, and eventually execution reaches 0081. (Return addresses are stored as 1 less than the address to return to.)
For the pipe glitch to work, you must be standing on the pipe end, and you must be ejected to the right (so that you cross into the next tile) while holding the down button on the same frame. By being ejected, your position no longer corresponds with the tile the game assumes you're standing on.
What criteria does the game use to determine whether Mario can enter a pipe?
I figured about the "end-cap" thing (SMB:LL does it for the World 5->8 warp after all) but what is this lazy calculation you mentioned?
It can't just be "Mario is on the edge of an end-cap tile" because while that works for the usual case, it also would let you enter a pipe by standing normally on its lip, which clearly doesn't work.
