Posts for ais523


Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Is it possible to get the analog stick to the position required on console using the "recentre stick" controller code, that changes what the controller considers to be zero?
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Wow, the trick starting at 21:30 in the encode is amazing (and I recommend around 21:36 in the encode for the screenshot, as a result, unless people think it spoils too much; the exact timing doesn't matter).
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I think that level-1 run is arguably submittable as a glitched low%, if you optimise it. (The game counts completed exits as an indication of progress, doesn't it? And you don't complete any exits.)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Is that actually the end of the game? Or did you decide to play only that part?
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
But what this paper is saying is that it isn't solvable, with current techniques, in a reasonable amount of time. (It's clearly possible to brute-force the fastest route; just that's far too slow.) In other words, this is a complexity problem, not a solvability problem. Incidentally, I was looking at the contents of the paper itself, and there are a couple factual problems: - Not all Metroid games (e.g. Metroid Prime 3) have situations in which you don't have the Springball, thus the proof breaks down there; - The proof for Pokémon-with-only-trainers falls down in the battle mechanics (not only is it possible to get an endless battle in every single generation, but in Generation I, the accuracy bug makes it hard to set up an unwinnable encounter, given that the player's team will eventually level up to 100). There's also a really simple proof that Ocarina of Time and Majora's Mask are NP-hard; you can encode the Travelling Salesman Problem in the games directly, using an event that kills you or makes the game unwinnable after a certain amount of time (think Fire Temple), a set of Small Keys with carefully calculated paths between them (which could be as simple or as convoluted as required to make them take a set amount of time to traverse), and a series of locked doors before the goal. The paper also seems to miss the result that Sokoban is PSPACE-complete, which allows for a very simple PSPACE-completeness proof for Pokémon Red/Blue (Strength puzzles are Sokoban-like apart from the victory condition, and Red/Blue's Victory Road has pressure plates that boulders can be placed on that let you match the victory condition too). This also made me start thinking about other games. A game with multiple sorts of keys-that-are-usable-once, locks-that-match-a-specific-sort-of-key-and-are-removed-by-it, and one-way corridors, is NP-hard by the same construction if it has at least 5 sorts of keys, and an inventory size of at least 2 (it could be unlimited). This allows for both pretty much all text adventure games, and also games along the lines of Pokémon Ranger.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Now, that's one of the better Sonic TASes I've seen recently. Yes vote, and love the category. (Is Knuckles capable of doing ultra-spindashes? If so, you should add something like "does not use ultra-spindashes" to the list of goals.) The glitched sequence in the final world was nicely reminiscent of unassisted glitched speedruns, actually, with all the amount of cautious-looking movement on platforms that weren't properly solid. TAS categories should require more of that sort of thing.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
It's probably most noticeable in Metroid Prime 3; the current SDA speedrun actually ends up going into hypermode during loading to avoid taking damage waiting for doors to open. (One of the effects of hypermode is that you're immune to damage unless you take enough of it to break you out of the mode; and if you don't attack during it, you get the energy tank you spent to enter hypermode back again when you leave it.)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I think the point is that skipping FMVs requires something equivalent to a crooked-cartridge trick, which is typically not allowed. So the category should really be "no rule breaking". And because that's implied, you can leave it off.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I'm pretty sure the rules allow you to use an unusual ROM if it has different tricks or glitches. A lower low% would be enough of a reason to use (E), I guess. Not sure if the run itself would be entertaining, though.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I thought 9% was impossible on Hard on (U) because there's a wall that you physically can't have enough missiles to burst through (and it respawns if you go back to get more missiles). (It's probable you two both know this too, but I thought I'd mention it for people who are trying to figure out the smileys.)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Good category, although maximum-rings might be a better one. My only problem is that if the run's as unoptimised as people are saying, it might be better to try again when you have time (or someone else can) and publish that. I'm not sure if it's better to publish the unoptimised run in the hope that someone will improve it, or not publish it in the hope that someone will improve it, though. (Either way, we can still watch the run, which is pretty entertaining.)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I was talking about the dash you did off the sixth target unlocking that door (which involved not shooting the target until the last moment).
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I don't use Dolphin, so I can't answer most of your questions, but: 1. The game can't tell the difference between holding down a key for two frames, and pressing it for a frame, releasing and repressing it between frames, and pressing it again, because it only reads input once per frame. So if you're trying to press as fast as possible in realtime, you're limited to pressing it once every other frame. If you're trying to press it as fast as possible in gametime, it might be possible to pause the game, release it and repress it, then unpause the game again, as the game would have time to read the release then. (Either the release or repress might need to be timed on the same frame, or the frame before or after, as the pause or unpause, depending on the game; you can experiment. Also, this trick doesn't work in every game) This is obviously slower in realtime, so it's limited to a few tricks, like increasing the maximum speed of a BLJ in Super Mario 64, where presses on successive frames are really needed. 4. Look for x, y, and z for position, pretty much every game engine I know stores them like that. You might want to try both ints and floats, and 16, 32, and 64 bit, until you find out the values. (I'd say floats are slightly more likely on modern consoles, like the ones Dolphin emulates, but it's far from certain.) Direction is more complicated; most likely, you'll have an x,y,z value for velocity, and some other value for facing, but there are a lot of possibilities. 5. Being able to use cheat codes is useful for testing, but the emulator should be able to emulate an Action Replay for you (I don't know how it's done specifically in Dolphin), so you don't need the actual hardware to do so. The main information you get out of this is whether a memory address is correct; you can lock a memory address to a given value with a cheat code and see how the game reacts to discover what the memory address does in the game. If you're trying to do a testrun of part of the game it's hard to get to (unlikely to be a problem with Super Monkey Ball), it's sometimes also helpful to use cheat codes for actual cheating, although obviously you can't do this for the final run. 6. Dolphin used to have a lot of sync problems, but many have been fixed by now. You might want to ask in the forum about Dolphin here about how much progress they've made with fixing sync problems. Make sure you use a recent version of the emulator (preferably, the most recent one that's known to work; it's probably hosted on this site), and if you want to guard against potential desync issues, keep frequent backups of your TAS (as well as replaying it from the start every now and then to ensure it still syncs).
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Haha, that's a neat combat dash before even entering the ship; it's surprising what sort of unexpected tricks come up absolutely everywhere when you TAS. What's the window for getting that like, I wonder?
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I don't see how it matters that the Wii version of OoT is implemented using an emulator, internally. It's an officially published game by Nintendo, who cares what code they used?
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Publish it; allow it to be obsoleted by a (J) run that's faster or a (U) run that's faster allowing for the difference in cutscene length. (In other words, only gameplay time should count towards completion time for determining obsoletion.) I think (J) is quite probably the better ROM for this (on the basis that the cutscenes are so long that not being able to understand them is a less bad problem than them being longer), but I think either should be fine. Murkier is the situation of, say, using 2 minutes of gameplay to set up a cutscene skip that saves 2.5 minutes in (U) and 1.5 minutes in (J), but I guess we'll get to that when it happens.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
To me, the obvious categories are 100% or MST, no BA/RBA any%, and glitched any%. "Partially glitched any%" isn't really a reasonable category, so just like with SML2, the glitched any% should obsolete the current any%.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
After reset-during-save glitches in earlier Pokémon games were widely exploited, saving was made pretty much completely bulletproof in Pokémon Ruby and Sapphire. If you reset during a save, the game will reload at the previous save, with a message that the save didn't complete correctly. (Presumably it does this using two save files internally, and a flag that records which of them was being saved to and whether the save completed.) (Then, they went and screwed up again in Emerald…)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
I didn't think the loading screens really got in the way. I enjoyed watching this, and think pure-speed is a viable goal for the game, but I'd have enjoyed watching it more if I knew what the object of the game was. Are you trying to get enough of the goals of each level that it allows you to continue?
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
That room just after you got the X-Ray Scope was hilarious. (I'm currently writing this during the reasonably boring period while you're doing the X-Ray Scope glitch; at least it's clear when it ends from the sound. It's a pity that can't go faster…)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
<dwangoAC> ais523: So 25 turns is long enough - if we get to Earth at 2012:37 and we have 25 turns (determined at the time of paralysis?) that'd be 2012 - 25 = 1987 <ais523> yep <ais523> I was born in 1987, now it's 2012, so on my birthday I'll be 25 <ais523> TASing mimics reality
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Hmm, I generally prefer to know what's actually going on in a run like this; the enjoyment isn't in seeing the encode and wtfing for me, it's in seeing a description of the glitches and routes involved. The submission message could do with a link to descriptions of the tricks that make this run possible.
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
While BrandonE was trying to publish something on a specific day, and the site wasn't working: <DrNach> okay this didn't work, onto plan B <sonikkustar> How serious is the problem? <natt> on a scale of one to steak sauce? <Canar> i'm guessing potato <natt> probably <DrNach> I feel like I'm playing Battletoads <BrandonE> 2 hours remain! :S <BrandonE> I wish this was like Battletoads...I can beat that game sometimes <DrNach> Onto plan C
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
If it saves time, do it. If it doesn't save time, don't do it on every stage. (How often would you be opening the menu for other reasons, resetting the glitch as a side effect, by the way?)
Editor, Experienced Forum User, Published Author, Player (44)
Joined: 7/11/2010
Posts: 1022
Mothrayas wrote:
Could also be something like World 1-2 of Lost Levels (VGMaps). It starts with a pretty large jump, especially if you don't know how to run.
At first, I thought it wasn't, but in retrospect, I'm reasonably sure that it was smb2j 1-2. (I remember the block just before the jump, now it's been mentioned.) Funny how games change with age… Thanks!