Discovering glitches through large-scale input bruteforcing is really infeasible. You would be much better off disassembling the game, understanding its logic and using your brain to abuse the weak parts and "create" glitches. I doubt bruteforcing would be faster no matter how powerful the computer was.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
Well, I'm afraid that would likely prevent things like the Excitebike phenomena occuring in the future, which I, for one, wouldn't want. If you remember, we had a run submitted (by FODA, iirc) and even published before Luke submitted his groundbreaking version. If FODA's run was submitted today without Luke's/JXQ's versions, it would likely be rejected as a bad game choice.
Basically, what I don't want is site-endorsed game limitations based on unstable opinions that can change depending on what one can possibly do with a game (which is yet unknown).
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
The TAS maker needs to know how popular the games to pick them judiciously. If you ask for opinions in the forum, you might get a few enthusiastic replies but that's not a good measure of how wanted the work is. Take out the guesswork.
Apparently, that's also dependent on how badly does the author want to see teh result themselves. Situations where an author would make a movie but not even submit it are numerous, and show no such disconnection. This suggests two completely different goals and motivation resources:
1) either you make a movie in order to get acknowledged and have a good movie as a consequence,
2) or you make a movie you want to see yourself and have people's acknowledgement of your effort as a consequence.
There are players who know the games they're TASing aren't popular, or even purposefully TAS unpopular games. They don't have this problem.
Basically, rating the games themselves might not be that bad a thing, but it can lead to unwanted results, such as unpopular games driven way further away from players (a very bad thing, actually), or, for instance, provoke some ego-related problems. I don't know.
If a player doesn't want to make a movie in the first place, why force oneself? If they do want, do they also need a warrant from the audience?
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
It's important to remember that "an optimal 10 second sequence will also be optimal in the first 5 seconds" isn't equal to "if you break the task into two segments 5 seconds each, together they will be equal to the optimal 10 second segment". In other words, if the first 5 second segment has a consequence in the second, the result will likely be suboptimal as a whole, which is pretty much the only reason there is little point in large-scale scripting as opposed to doing most of it by hand.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
That's because a 10s segment is far too large. What can be played by hand will have to be played by hand anyway, and inputs that have no effect on the outcome don't even have to be included in the tests.
So, for instance, a 10 frame long luck manipulation segment where only left, right, down, A, and B make any difference, will take just a couple hours or so. For longer segments involving something other than that, heuristic algorithms that could consider usefulness of certain input will be necessary (however, if one is able to script such a bot, that would surely be enough to test the required action by hand, anyway).
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
My money's on conventional computing getting obscenely good at parallel processing before then anyways. Someone's gotta be working on computers with one or two powerful processors and like 20 simpler processors by now - properly programmed that sort of thing could be monstrously fast and still reasonable to build, I think.
It's actually closer than you think. :)
The contemporary GPU technologies already allow cramming not 20, but hundreds of such "mini-cores" into one GPU, and you can have several GPUs in one PC as well. Considering that bruteforcing is hilariously parallelisable, maybe we can still reap some usefulness from that kind of approach.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
I don't think you can get more "ungrammatical" than "different of".
Also, "ungrammatical" sounds very… ungrammatical. That is, grammatically incorrect.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
Except for SGB. SGB should have a 4/3 aspect ratio.
It is true that it's displayed in 4:3 on a TV screen. However, the problem I see with that is that the Gameboy screen in SGB encodes still appears stretched in comparison to the normal GB screen image. Considering that GB is the original platform of these games, it should take precedence in that particular respect (i.e., why should we purposefully distort the image when there's no real reason to?).
[EDIT] Elaboration.
We encode GB movies as SGB so that they look better. Distorting the image doesn't look better. If the viewer still wants to distort it, there's the option to force a certain aspect ratio in many popular media players.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
"glitchless" sounds less arbitrary then what would be needed to make the current runs qualify. The problem is you would need to define a list of allowable glitches, and a list of banned ones.
The problem with "glitchedness" is that it's already hard to define what glitch is. Is data manipulation a glitch? Is abusing collision detection algorithms' "defense mechanism" (zipping, ejecting) a glitch? What about abusing the character's motion physics to produce unlikely in-game situations? On the contrary, "going out-of-bounds" has a rather clear definition: colliding with the terrain in a way that makes you go through it. It should be quite enough.
Satoryu wrote:
it's not an any% by definition, as it collects the lowest possible amount of items.
Actually, it's rather the other way around: collecting the lowest amount of items is a consequence of beating the game as fast as possible, because neither of them except these six save time. As you said earlier, the existence of low% as a category only makes sense if it's not the same as any%.
Saturn wrote:
It's really surprising how many people gave this NBMB run a sub-9 technical rating, despite that it's the by far most optimized Super Metroid TAS we have here, with all of the new techs and long-time worked out optimizations included.
What is most surprising, of course, is that you give 9 in technical rating to all Super Metroid runs you see. Let's name them: any%v1, any%v2, ingame any%, 100%v1, 100%v2, NBMB.
Does this mean they have all been made on the same technical level, consciously or not? Or does this mean anything else, particularly the fact that you've always tried to hide the majority of improvements that you knew full well since long time ago, and praised the authors for their always "most optimized" runs without pointing out how improvable they actually were until it was too late?
I think there is a word for this, hmm… let me remember it… starts with "hypo" and ends with "crisy". I think this word is the exact reason I don't really want to have any further communication with you, ever. Don't even waste mine and others' time.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
When you come off your high horse (and only then), be sure to go and check how much time it takes to get into the wall and climb it somewhere where it's actually useful. So far only one such place has been confirmed: the gravity suit room break-in.
This particular glitch has been known for ages, mind you.
The MB skip warp requires a very elaborate setup I described up there, there is another way to trigger it, but it's even longer. I mean, sure, test it.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
Is low%, real time the same as any% real time with this glitch, or can items be further skipped costing time?
No, getting additional items will not speed this run up, nor is it possible to skip more at the moment.
Kirkq wrote:
Does it seem like going out of bounds will speed anything up in the 100% run?
Might be possible for in-game, but likely not for realtime (I don't really see how several additional room transitions + death cutscene + saving/loading cutscenes + intro screen + x-raying up the wall will turn out any faster than just fighting MB with all beams).
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
Cpadolf, basically, the only criterion you need here is that no other runs go out-of-bounds. Every single glitch in normal runs is performed within the level space, conforming to the room bounds.
Glitched Zeldas are all OOB. Same goes to several NES Megaman games, although it's notable that the non-OOB versions of their TASes just didn't exist.
Saturn wrote:
So is that the reason why you don't work on your low% despite sitting around here every day doing nothing for years?
No. The previous time I explained my reasons was the previous time, go find it yourself if you want.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
The problem is when you have about 20 catagories for a game, is that its inevitable you'll get a gross inconsistency in technical quality between runs. Often a single discovery made during one run can suddenly make them all obselete this is partly the reason for keeping brances per game minimal.
It might be a good choice overall, but you have to satisfy people who want more of one game if that game allows so. 70-star SM64, SMB3 full amazing run™, SMW small-only, etc., all fall under the "not enough while there could be" type of demand.
As for Super Metroid per se, it's known to become obsolete already during the production of a run due to the mad pace of discovering new esoteric tricks, the players have already yielded up to that.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
