Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
In terms of input button mapping, the DS and SNES have the same buttons if you don't count the DS's microphone, lid switch, or touchscreen. (And there are plenty of DS games that don't really use any of those.)
Probably the best rule for runs designed to synch on two entirely different systems would be to give another input every time the system in question polled input (normally, but not always, once per frame, sometimes not counting lag frames). Presumably you'd have to scale the speed of the videos to match so that the similarity in input became obvious.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Patashu wrote:
What about a Super Hard max score run on each level? (If I'm correct in remembering there are only finite resources in each level)
Indeed there are, and I'd enjoy watching that too. However, the killing of every non-respawned enemy, etc., that's needed to max out the score would make the movie take a while, and routeplanning might be a pain as a result, not to mention actually making it. And the number of people with lower tolerances for long movies than me might mean it wouldn't even be accepted, which would be a shame.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
FractalFusion wrote:
The scrolling Japanese text isn't supposed to be an issue here. No one can read it anyway since it scrolls so fast, and understanding it won't improve anyone's opinion of the run. The fact is that it's Go, and it's not very popular around here. If it was chess, with strange and counterintuitive moves from both player and AI, the reception would be much different.
I thought the reception for chess games in TASvideos was normally awful (sufficiently so to become a running joke/meme at one point). I suppose the fact that I'm aware of the rules of Go and how a game typically looks (although I'm not very good at it myself...) makes the run look a lot more impressive to me. In particular, the strategies used by the player all through the run don't make a lot of sense (not just playing in the corner); the AI plays far too passively, and ends up losing by a larger margin than is easily possible, which to me is a lot more impressive than the chess TASes where the AI lost due to tactics rather than strategy.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Wow the AI screwed up on that one. Pretty impressive as far as it goes, although I was left wondering "it's over already?". (I ignored the cutscenes due to my inability to read Japanese, although those aren't the main draw of a TAS anyway.)
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
The order in which the levels are done makes this even better; it just gets more and more insane as time goes on. (A special call-out goes to Doodle Woods, for doing the level in completely the wrong order.) Well done at making the Golden Passage not look completely boring, too.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Jungon wrote:
Voting Meh here =P
Look, I know you got the drawing skills, but this version just isn't as entertaining as the old (and awarded) one, .. even with the "different ending", I think (of course not considering a No vote) that if you want to do an Art Academy version, you'll have to get a 100 new images to draw, because seeing the same as before was just less than expected ...
I disagree with this opinion (haven't voted yet because I haven't watched it yet, btw). In tool-assisted speedruns, you often get obsoletions improving routes, improving optimisation, using new tricks to gain a few frames, even though much of the input nevertheless stays the same. This is the same idea for an entertainment-based tool-assisted superplay; even if it isn't the entertainment equivalent of an entirely new route (which might well make the finished TAS less entertaining...), it just makes incremental improvements and leads to a better product overall. Really, "has it been done before" isn't the right question to ask; rather, "is it the best version of this run". When I watch the run, I'll be voting on whether the art as a whole is more or less entertaining than the previous version.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
I always thought the technique used to end input was part of the category of the run. That is, "fastest to end input" and "fastest to game completion" are two different categories (the first obviously having a shorter time via TASvideos timing, but the second, which isn't allowed to end input until after the game is completed, being more interesting for some games). Obviously, you wouldn't publish both (unless for some reason they had entirely different routes; wasn't there an April Fool's TAS that completed half a game with basically no input at all, just luck manipulation so that the first half of the game would complete itself from there?), but you'd choose whichever category was better.
So all that really needs doing is to add a category explaining the issue, and the run suddenly isn't suboptimal any more.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
There are multiple distinguishable categories here: "glitched any%" (the published run); "all golden coins" (which would allow any glitch to get to the coins themselves, and allow skipping of levels like Tree 3 and most of Macro Land); "all levels" (either exit on all levels, although some levels would have to be done twice in order to unlock other levels); "all exits". Additionally, the last two could allow either all glitches, or only a subset. I'm not sure if all levels and all exits would be substantially different, or even different at all (is there any level with multiple exits, where one of the exits doesn't unlock anything?).
As for which category to go for, I'd suggest all levels; there aren't that many secret levels, after all. I'd be inclined to allow all glitches but the pipe glitch (as the glitch that sets it apart from a large-skip game), together with any other glitch that allows Mario to go beyond the bounds of the level as a whole (going through a wall should be fair game, if there's a glitch that does that). The reasoning here is that there aren't very many bonus levels compared to the length of the game itself, and several of them are fun to watch (there's one autoscroller on the main map whose only purpose appears to be to get large combos for entertainment purposes). I also have no issues with the pixel tricks (either old or new), but perhaps that's because I like glitches which demonstrate those sorts of understanding of game mechanics; the only real annoyance that it causes is the chaotic behaviour of Mario's sprite.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
If any% is very similar to 100% (which seems likely), do 100%. Any% would only really make sense if it's possible to get a much lower percentage than intended, e.g. if you can glitch past note doors somehow.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
This was great fun to watch, the category issue's the only thing that might potentially hold it back from publication (although I don't see why we don't have as many categories for a run as people are willing to watch...) This should at least go in Gruefood Delight if it's rejected (which seems likely atm).
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Bhezt Rhy wrote:
Well, I am TASing Pokemon Ranger in Nintendo DS. Maybe this colud be interesting......
I'd rather like to see that, actually, even though it'll mostly be boring waiting sequences between drawing lots of circles (or triangles?) as quickly as the game can follow.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
errror1 wrote:
so you found a glitch that lets you kill yourself faster
That definitely needs to be a run goal.
"Abuses game code to die quicker"
Well, we're just resurrecting an action before we died, which is obviously a useful thing to do. Annoyingly, it seems to be incompatible with the "throw" action; it /is/ compatible with the "cast spell" action, though, which means it's still likely useful on Water. I just need to figure out exactly how all the timings work out there...
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
errror1 wrote:
I was able to follow the last encode fine watching it frame by frame You murderer!
How do you eat the brain of a ghost?
I'm not sure, but the game lets us, and it saves time, so we do.
Meanwhile, I think I've discovered yet another new glitch, this time in command repeat. Try this in debug mode (it works outside debug mode too, but is rather harder to set up for the test): wish for two amulets of life saving and a wand of death, put on one of the amulets, zap the wand at yourself; as expected, it takes a turn, and you die (with the amulet resurrecting yourself), and you get the usual "You survived that attempt on your life" message. Now, put the other amulet on, and this time, instead of zapping the wand at yourself, use command repeat to zap it at yourself three times (i.e. 3zw. with vi-keys controls, or n3zw. with numpad controls, where w is the letter of the wand). You end up dying and resurrecting from the amulet, but you only get half the usual cutscene; the lifesaving doesn't take a turn like it normally does, and the zap doesn't take an action, like it normally does. This seems a pretty useful discovery, as it's possible to cancel out both the time penalty for lifesaving and the action before to do something in no time at all, which would be pretty amazing, although it can only be used in situations where we need to die anyway (because otherwise, the time saved is cancelled out by the time lost putting on the amulet). Some notes about the glitch: it isn't compatible with using life saving to cancel out the time penalty of something else (as doing so overwrites the command repeat variable), so it's probably useful once in the run at most (for safe-teleds deathwarps, which AFAICT are useful only on the plane of Water, and only once); and I'm still not sure I fully understand it, because although I have a basic idea of what's causing it (http://nethackwiki.com/wiki/Source:End.c#line428 and the line after are the culprits), I'm not sure of the exact details, and haven't managed to reproduce it with more complicated deaths than the very trivial suicide explained above. Still, I'm hopeful.
(Incidentally, the other half of the lifesaving "cutscene" - the "you survived that attempt on your life" message, can indeed happen, at the end of an entirely unrelated helpfulness; I've seen it happen in someone else's unassisted game, but nobody was sure what had caused it at the time.)
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Tafatt wrote:
Was not this submission rejected because of the glitch this movie is using?
It was rejected because the emulator didn't emulate the glitch correctly, which obviously makes the run impossible on a real console. The emulator has been fixed since, so it now emulates the glitch correctly; it turns out it's still possible, but slightly different. So runs using this glitch with a recent version of the emulator are allowable, and runs using the old broken emulation aren't.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
rhebus wrote:
ais523 wrote:
(It turns out there was no nearby seed that let us bring all five wraiths down nearby, so we get four of them, and one of them gets away. We might nab it later, or gain the 14th level via polyself rather than wraith; there are plenty of options.)
I thought that wraiths can still leave corpses on graveyard levels such as the Valley, it's just 1/3 as likely. You can kill and eat that last wraith in the valley for the 14th level.
We did kill and eat them on a graveyard level. The issue is manipulating a corpse at the same time as manipulating misses from all the wraiths surrounding us; diving to a level with a maze layout made it much harder for the wraiths to hit us. This bit annoys me a bit as it's a little sloppy, but it seemed better to take a "good enough" result rather than waiting over a second for a perfect seed...
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Another WIP encode is available (at http://sonic.net/~ac/tas/nethack_tas_20110103.mkv encoded by Ilari; as usual, that account has a relatively small cap, so although the video is small it would be nice if someone rehosted it with more bandwidth). What I've done since the last one is to get hold of a magic lamp (used) and a magic marker (not yet used, and left pretty much where I found it), fixed god anger using a dead shopkeeper, started training detect monsters (so you all can see what's going on), then went back to the Valley, then to the top level of the Wizard of Yendor's tower, to gain the levels required for the Quest unlock. (It turns out there was no nearby seed that let us bring all five wraiths down nearby, so we get four of them, and one of them gets away. We might nab it later, or gain the 14th level via polyself rather than wraith; there are plenty of options.) We've saved prior to the first actual sequence break in the game (as opposed to "intended" routes through the game which are far too risky to do without, at the least, strategies designed specifically to survive them; the route through the Castle that we've taken so far can actually be done in realtime, but only with around a 1% success chance based on Maud's records). What I'm going to try to do is use the master mind flayer's ranged attack, which doesn't require line of sight (unlike pretty much every attack in the game), to wake the Wizard of Yendor from outside his tower, then rely on the cheating AI to take him through the nearby impenetrable wall to fight us. (This can be done in realtime, but it's a bit random; I'll see if I can figure out how to do it quickly.)
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Mister Epic wrote:
Just the fact I saw the Kururin (if it's called like that) getting really, but REALLY, near obstacles and the records being destroyed makes me want to vote Yes.
Kururin's the main character. In English, at least, his vehicle's called the Helirin. (I assume it's the same one between the original game and the sequel.)
Also, a pretty impressive improvement! Some of the changes in the minigames were immediately noticeable, and it pretty clearly needs TAS tools to pull off some of those shortcuts consistently.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
eternaljwh wrote:
Would a lowercase one, a la
.....#
####.#
.....#
.###.#.
.......
be valid to generate? (Dark spaces won't matter)
That pattern also has a pillar. The maze generator never generates a loop you can walk around, like the top of the capital A you drew earlier, or the bottom of that lowercase a.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
eternaljwh wrote:
ais523 wrote:
Oh! I know what you can do to be entertaining. Manipulate one of the dungeon levels to generate looking like a word. (Probably use a MMap scroll on it.)
Is that even possible? I can't think of any dungeon generation algorithms that would produce anything looking even approximately like a word,
T A S should be easy enough. Branching corridor, corridor up to room on each side with corridor below room, winding corridor+room. Or, in maze,
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
eternaljwh wrote:
you seem to be fixated on AC, which is relatively irrelevant
Probably because I'm in that "can ascend sometimes" bracket; AC is more important than HP in normal runs pretty much the whole game, since it reduces the damage you take from attacks as well as preventing you getting hit.
Probably because I use E only rarely.
So:
How many points is Demogorgon worth? I wonder if it's possible to hit MAXINT score before turn 2000. Probably not.
Demogorgon is insignificant in this. UncleBob/berry (who came up with the infinite carry capacity glitch) also came up with a way to get maximum score without wasting time, though, although it would happen after turn 2000, and possibly require an annoying amount of hexing. There are actually quite a lot of things you can do that don't cost a turn, and some of them reduce score; we can easily go over MAXINT before turn 2000, and reduce the score at the altar. (Staying below would mess up the infinite carry capacity glitch, but messing it up in zero time at the last moment just before offering the Amulet wouldn't hurt anything, as you can offer the Amulet in zero time even if completely overloaded.)
Anyway, doing something normally 'unwise' on the penultimate move (if it doesn't make the turn counter advance) would be interesting, at least if it dumped like NAO games did. Angering a god (and surviving) to get the "I believe it not!" message on the high altar would make for a good screenshot at least.
The official version of NetHack doesn't have a dumplog feature (although the version shipped with Debian does, strangely; it's one of its few changes from the official version). Thanks to the magic of TAS tools, though, I could easily create an official-looking dumplog for the game (in fact, a plain ttyrec would probably be enough).
If it were even remotely possible I'd suggest ascending with one or more woodchuck pets; however I do not think it is.
It may be (cursed figurine autotransformation + lots of luck manip), but I think it depends on the turn/action correspondence on the last turn.
Oh! I know what you can do to be entertaining. Manipulate one of the dungeon levels to generate looking like a word. (Probably use a MMap scroll on it.)
Is that even possible? I can't think of any dungeon generation algorithms that would produce anything looking even approximately like a word, apart from maybe the "caverns" scheme, which would still only be a relatively poor approximation, and require probably years of luck manipulation. Making monsters line up to form a word makes more sense, although it's been done in realtime before now; however, we'll have to collect everything needed to do that anyway apart from appropriate statues, and it would only be a few turns out of the way to collect the statues.
In other news, the strategy I was planning to use for Air doesn't work, but I think I've found an alternative. You can reduce the helplessness from hurtling to the rest of the turn (as opposed to the rest of the turn, plus 8 more turns, the usual length) by lifesaving; unfortunately, if this is done via a regular death, it'd set us back to natural form (2 actions per turn), and obviously air-E form and its 4 actions per turn would be better, as the spare turns could at least be used for walking. We can get one instadeath on Air by setting up food poisoning at the right moment before T:2000 (the timeout really is that long!), but only one, and I need to check whether it happens just before or just after the start of a turn (which will make a huge difference to the timing, due to the loss of the rest of the turn). Other potential strategies all involve becoming helpless via two different means at once, then cancelling one in order to simultaneously cancel the other (due to a glitch); being attacked while asleep (and not in the first turn of sleep) can reduce remaining the remaining helplessness timeout to 1 turn, but that seems to always cost 2 turns and is always useless; theoretically the fastest might be to reduce the helpless timeout to 0 by getting a nymph to steal a cloak while simultaneously helpless from hurtling and helpless from fainting from hunger and immediately after steal an unrelated item, but I'm not entirely sure if that's even possible to set up (I failed in wizmode, at least), and it relies a lot on the timing of all sorts of things; if it is possible to set up, it's probably only possible to set up once on the whole Plane.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Bisqwit wrote:
Something I recommend for entertainment:
Make every prompt and action display clearly for at least 1 frame.
People who watch movies frame by frame, trying to follow what's going on, will thank you for doing that.
I mean, if you're expending time doing some gimmicks, you might as well make it worth watching. Otherwise, you're better off writing fanfiction than making a TAS.
I agree with that. I was thinking about submitting two movie files, one "as fast as possible" for people who like to not know what's going on, one slowed down so you could actually see what's happening. Ilari was thinking about hacking JPC-RR to do screen redraws in the middle of frames to allow a hacked encode that showed what was going on at the subframe level, which comes to much the same thing. There'll definitely be some way to watch something played at humanly plausible speeds but with TAS strategy and luck manip, though.
Editor, Experienced Forum User, Published Author, Player
(44)
Joined: 7/11/2010
Posts: 1022
Patashu wrote:
Kill or make extinct some unlikely monsters, like djinns, skeletons, player roles.
http://www.steelypips.org/nethack/extinctionist-faq.html
The method listed for killing as many mail daemons as possible is especially ingenious. Could it be improved through manipulation of monsters?
Bring an item to a ridiculous positive or negative enchantment level.
Solve sokoban in a ridiculously rule-breaking way. Maybe arrange the boulders into letters/smiley faces then cheat your way through?
Can you get engulfed by an air elemental and have it bring you places at high speeds? Would it ever be useful to get an extra movement around?
Ring/amulet devouring fest?
Get some ludicrous number of useful pets, flood a level with them?
Get so much AC that it rolls over and becomes negative?
Get surrounded by (insert big nasties here) and not die?
It's a pity the aim is as fast as possible, else you could have fun with taming/removing the corpses of the riders for good.
Grappling hook! All the cool kids on the block have one. Also magical instruments and royal jelly.
Amusingly, we're planning to do (or have done) some of those. We've had a ring-eating fest already, for instance (which has pretty much lead to me to refuse to use xorn form in this run ever again...), it turned out that hurtle-cancelling doesn't work (I misread the source) so we may have to move while engulfed by a hasted air elemental to do Air (as jumping and throwing iron balls don't work there either), and we need to flood Rodney's with monsters for the enexto glitch, and they may as well be pets if you think that would be funnier. Also, we have enough HP at this point to habitually survive just about anything but an instadeath, and instadeaths can be manipulated away.