I think if you are going for something like all gold crowns or all max scores, no-clip is redundant (because by definition you can't use it to skip important parts of the stage where you need to get points!)
Sorry to complain, but what is taking so long with this movie being published?
Probably making the official encode.
Gamecube and Wii games are notoriously hard to make encodes of due to Dolphin's a/v sync and avi dumping issues, for example often leaving gaps in audio that have to be fixed.
In my opinion, a 100% (or "All Items" as the author calls it) run for Zelda should at least include beating the game with all heart containers, a full inventory and maximum upgrades for persistent items. It's arguable whether maps, compasses, quantity upgrades and red water of life should be required or not, so I'm leaving them out of the discussion.
Still, based on my requirements, this run doesn't qualify. The problem is that there are no bombs left in the inventory at the end of the game, creating a hole which can clearly be seen in the sub-screen when the TAS switches to the bow upon entering Ganon's room. Unlike arrows which don't disappear from the inventory when running out of rupees, keeping at least one bomb is required to have a full inventory.
For those that would argue that picking bombs once is enough to qualify (a position I disagree with but has some merit), well in this case the run is sub-optimal. The same logic has to apply to the bait (1 use), and yet it was purchased twice. This is inconsistent, and I don't see any arguments that would reconcile "no bombs left" and "detour to purchase extra bait" for 100%.
Therefore, this submission should be either fixed or rejected in my opinion.
Oh no, it's the last submission of 'all items' all over again :D
Would something like "no memory corruption" or "no memory overflow" be a good description for the old run?
It abstains from two things:
The 'coin case arbitrary execution' route (executes data as code, corrupts memory via controlled writes)
The 'box cloning/bad clone/map distortion' route (resets mid-save, corrupts memory/buffer overflows a lot)
If you call controlled writes to memory memory corruption (and I think this is correct) then 'no memory corruption' should work on its own.
Fun fact - the TAS for this game was uploaded to nicovideo about a day later than the one for OoT, but it passed 100k views before the OoT TAS did (though not by a huge margin) :D
You might recall that Hero Core, by the same dev, has 'Reallyjoel's Dad' difficulty which is possible only for a TAS:
Link to video
Hyper Princess Pitch also has a similarly absurdly tough secret difficulty, 'Reallyjoel's Mom'. It looks like it would be a challenge to try and beat in TAS.
Link to video
What improvement will be found for OoT any% next? Ocarina items without a bottle? A variant of Down+A that you can do in Gohma's room? Death Hole Wrong Warp in non-3D OoT? ;)
This glitches all battles to end when you try to send out a Pokémon, and for the case of trainers the game assumes you have defeated them!
This is not a glitch - it is deliberately programmed into the game to end the battle if it notices you have no alive pokemon. GETTING to this state is a glitch, but what happens is not. (There's also a similar failsafe where if you haven't gotten any pokemon yet, the battle won't even start. This was used in the last submission by TheZZAZZGlitch.)
Finished watching - yes vote. Thanks a lot for making this! Having a TASed example of the RTA route pushed to perfection should make it easier for glitchcity guys like MrWint and TheZZAZZGlitch to study and maybe find improvements on the map distortion techniques used :) And of course this will be great to show to people and see their reactions :D Lovely disco colours after the elite four. Too bad they couldn't go on even longer :)
(Is anyone planning on uploading this to nicovideo?)
This is probably the biggest part of Pokemon TASes, as luck influences Pokemon starting stats, criticals, what Pokémon goes first if speed is tied, the damage range of a hit, if they hit or not, wild encounters, etc.. For the most part luck can be manipulated by delaying input for a few frames at very specific times. Wild encounters can sometimes be manipulated based on your movement patterns.
I thought that gen 2 used the same RNG as gen 1. That is, every single thing related to input influences the RNG unpredictably (pressing buttons, releasing buttons, timing, etc) which is how MrWint could make a bot to optimally TAS pokemon blue by trying different input until it got the luck it wanted.
I think that if the bad clone/map distort route is faster, it will be the accepted TAS.
And if it is slower, then they should be allowed to duke it out (speed vs entertainment, what is more important?)
(It will also encourage further exploration of how the glitches work, to have a well optimized TAS showing it off.)
Therefore, I would submit it.
Lmao this hack got wrecked. I'm glad you could only horizontally underflow once because otherwise not enough of the game would get shown off. I think that doing a normal+camhack side by side encode would be the best way to show off this TAS (kind of like the sonic 2 submission) as both where the camera is and where sonic is are of interest esp. in the third stage.
