Posts for Nightwatch


Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
IMHO, annoying and confusing camera angles aren't worth 3 frames. Voted Meh. If it were much faster than the current one, I'd vote yes, but 3 frames don't cut it.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
DarkKobold wrote:
I feel like this video deserves a tag line in the publication, like the two 14% runs of Super Metroid, but it isn't sufficiently different or interesting to the non-punchout enthusiast to warrant a separate publication.
Good idea. This isn't a new speedrun, just an "alternate run" to the currently published one.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
I don't post much here, so take my opinion with a grain of salt, but... It seems that one obvious thing to do to solve this problem would be to have an official section for shout-outs, thanks, and credits for each run. A lot of people do that informally, but it'd be great to have a rule become somewhat official: if you use techniques or input from someone else's run or WIPs, then you should credit them with the techniques and input, and include a link to the run or WIP. That way, credit goes where credit is due. Kinda like scientific citations, intentionally not doing this would be considered plagiarism. Also like scientific citations, people who don't currently have the WR run on the site but came up with lots of techniques and tricks would get recognized.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Easy yes vote. I loved the autoscrolling madness.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Looks like we could use a FAQ about configuration settings in DeSmuME to avoid this happening in the future...
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
A solid run that's entertaining and well-executed, with clear, simple goals. Easy yes vote in my book.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Looks solid and kept me entertained. Voted yes. I definitely endorse the subtitles, by the way. Subtitles are much more user-friendly (for the casual watcher) than a long scene-by-scene submission description.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Looks pretty good. It's not the most entertaining game in the world to TAS, because of the slow speed, but it's good enough, and technically it looks solid. Yes vote.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Desty wrote:
I guess the question is whether it's possible to do a fast diff of each state against the last, especially on larger states like on the N64. Maybe some hooks could be added to the emulation code that add to a delta record for the current frame during memory/register writes?
It'd be interesting to see an analysis of how much (as a ratio) of the state of the machine that a typical game alters every frame. I'm curious: I have no idea what the average number would be.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Would it be possible to use it for at least a few laps though? As long as you gain back enough rank to make it to 3rd by the end you should be fine. Besides, it's more entertaining to not use it for every lap.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Maybe the Q-Gears team might be able to help shed some light on what's going on? Given that they have the FF7 script system figured out and boss battles repeating sounds like a script bug, it might be worth a try...
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
TheRandomPie_IV wrote:
Another small addition that would be nice is a function to load config files, as there doesn't seem to be a(n easy?) way to do this at the monent.
The "Lua way" to load config files is dofile(), which is part of the Lua standard library. Just write your config files in Lua and load them at runtime.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
moozooh wrote:
gia wrote:
Long tas? there's way longer, long boring tas? pokemon gold is 2 hours plus, pokemon is boring, easier to watch in fast forward.
Using Pokemon as an acceptance backdoor for every TAS that's also very long and boring doesn't sound convincing. For one, we also have Jaws published on the site. Popular opinion indicates that nearly every game is better than Jaws. Ergo, we should publish every game?
To me the biggest problem, which is stated in the description, is:
The reason for the low re-record count is simply because this game is extremely easy to TAS. Once you have the strategy for a level planned out, you don't need many re-records at all to record it.
This is what makes this run different from Pokemon, which is more difficult to TAS because of all the luck manipulation. Things that make TAS's difficult also tend to make them more entertaining (with a few exceptions, such as run speed optimization in SMW). If Pokemon had no luck manipulation potential, then it'd be just as boring. The fun of otherwise boring TAS's like Pokemon is the way a TASer will win despite insurmountable odds by cracking the RNG. But this run looks like a video version of a GameFAQs walkthrough. Edit: Donkey Kong for the NES is a good example of what I'm talking about, I think. Originally, it was a below average TAS, good but not particularly interesting. Once the "climb down the up ladder" trick was discovered, though, it became awesome. That's the kind of thing a game like this would need to be TASable, I'd think.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Warp wrote:
Are you really talking about the same thing? Knowing how iPhone works (basically it has no buttons, just the touchscreen) I assume the graphical buttons double for real buttons which you use to play. It would feel somehow logical that replaying a movie also "presses" those same buttons (if that's what you mean).
Well, what I mean is that it displays a photo of a NES controller in the bottom part of the screen, which I assume was similar to what you were suggesting: a photo of a DS. Photos don't really work well as user interfaces and for emulators they tend to clash with the sharpness of the game display. If you were thinking about a little diagram or digital art of the controller, then I'd be all for that. That might even be useful for other videos too: it's easier to read a diagram than a "LDUR BA Ss" status line. (Of course, then you can't do JXQ-style art. Oh well...)
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Warp wrote:
Blublu wrote:
Hey, that gives me an idea. If the "DS" is visible, you could show the input on it, too. So when A is being pressed in the movie, the A on the DS would be pressed as well. And so on. It would probably be a bitch to implement, though. But it would make the movie extra 1337. :)
That's actually similar to the idea I wrote in the DS group some time ago. I like and fully endorse your idea. :)
I dunno. Some emulators (like the iPhone NES emulator) do this and I always found it kinda cheesy and distracting.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Tub wrote:
so yeah, worst case is: you have to re-create the whole rendering engine in LUA, using reverse engineered ram values :/
Yeah, this is probably what would have to happen for that to work, unfortunately. You're basically talking about changing the screen resolution of one of the screens. For some games, it might be more feasible to move the top screen farther down (i.e. eliminate the gap). If the game is actually using one playfield in memory and rendering it from two camera offsets, then it might be relatively straightforward to change the offset of the top (or bottom, or both).
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Voted no. This game just isn't TAS-able unless some major glitch is found. Like LagDotCom said, if I didn't know this was a TAS I'd think it was a video walkthrough.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Floogal wrote:
I think that the washing machine form is more than just a cosmetic change. I recall going around Bubblegoop Swamp, being unaffected by the piranha waters, but also being unable to climb the tall, thin plants.
I think those are just bugs that Rare didn't have time to fix. When you turn into a washing machine, Mumbo says something along the lines of "This is black magic, don't go complaining to Mumbo if not everything works right".
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Is it just me or does this violate the "end the movie when there exists no input that can end the game any faster" rule? Either way, I'm voting yes, but maybe that last part should be redone (it's the end of the movie so it could be easily changed)...
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
The cheat system in BK is actually pretty horrendously complex. I guess I might as well put the info I've collected down here, since I've never seen it all gathered in one place before (and there's lots of misinformation). There are *five* types of cheats in BK, arranged from least to most egregious (at least in terms of what the game apparently thinks): (1) Cheato upgrades: obtained from Cheato in three hidden places in Grunty's Lair, these upgrade your max item counts. Only usable once unlocked. The effects are permanent and are saved in your game file. (2) Bottles' bonuses: obtained from Bottles minigames after completing the game, these cheats change your character model. Only usable once unlocked. The effects can be reversed by entering NOBONUS or by restarting. Only the unlocked/locked status of these cheats, not the effects, are saved. (3) "Regular" cheats: give you unlimited health, air, honeycombs, etc. These were known while the game was in stores. I assume Rare put them up on their web site or spilled them to magazines. You can enter these as much as you like, but the effects aren't saved in your file. (4) Stop 'n' Swop cheats: unlock the eggs and the ice key. Rare never revealed these: most likely they were going to be revealed in Banjo-Tooie. There's no penalty for entering these. The effects are permanent and are saved in your file. (5) Sequence breaking/"strong" cheats: let you sequence break by modifying parts of Grunty's Lair. Rare never revealed these: they were found by disassembling the game. Using one of these cheats bumps the cheat counter, which is not saved in your file (so you can save and quit to reset it). When your cheat counter hits 2, you get the message "Stop this cheating, Grunty says, or your Game Pak I'll erase!" If you then attempt to enter a third cheat, Bottles will prompt you to confirm that you really want to enter the cheat: if you say yes, the game will erase your file! Because the state of Grunty's Lair is saved in your file, the effects of these cheats are saved too.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
Bag of Magic Food wrote:
It's too bad the "legitimate" cheats would never save time. Now I wish they were part of the 100% requirement. Though I guess the home puzzle ones that alter your character aren't even saved to your file, are they?
None of the cheats except the mild ones that Cheato gives you are saved. The game doesn't even save the "strong cheat" counter (the one that deletes your saved game if you use too many "strong cheats"; i.e. anything that alters Grunty's Lair). So you can enter as many as you want without penalty if you save and quit after every two cheats entered.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
rabid squirrel wrote:
Ok, this is Mr. I-just-discovered-TAS-and-I've-been-watching-TASes-of-all-my-favorite-games-on-youtube-in-the-last-three-days speaking (meaning I have no experience at all) I watched the Banjo-Kazooie run, and I loved it, but I'm hoping somebody does an any% run. IMHO, an any% run gets to the essence of what is possible to beat the game in, and I would think that's the meaning of life, the universe, and everything! (at least in the world of TAS :P) Just my two cents, and a bump for this thread in the hope of reviving it (and maybe convincing bkDJ to continue).
Banjo-Kazooie has such a tight requirement on the number of jiggies needed to complete the game that an any% run would be very similar to a 100% run.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
This hack is far more technically impressive than even Super Demo World: TLC. I'm a little skeptical that the author of this hack saved any time hacking SMB1 instead of just writing a new game from scratch! Honestly, if we allow any hacks at all then we should allow this one. Yes vote.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
The only option I'm really opposed to is (3) - it's too arbitrary. Personally I'd go with (1), because I don't really find the wobbling that distracting: it doesn't apply during most boss fights, in autoscrolling segments, and while Yoshi is transformed anyhow.
Experienced Forum User
Joined: 2/12/2008
Posts: 67
Location: San Francisco Bay Area, CA
DarkKobold wrote:
Unless the game is capable of saving in a single frame, which may be unfortunately often... Which sucks, because I love these save corruption runs. Also, console vs. TAS save exploits will be enormously different - TASes can only reset consoles at the frame marker - a console can technically stop the code at ANY point during the save, opening tons of new possible exploits. Unless, of course, the coders of the emulators allow for resetting the game at any point in code execution - but I doubt any coders would, seeing as the trade-off in time to implement versus the pay-off in TASes wouldn't be worth it.
Hmm, I hadn't thought about it that way before. I guess what it boils down to is that movie files are based on the assumption that input can never affect the game until the game decides to read it (i.e. no interrupts are generated on button presses), and that the game reads the input at most once per frame. These are pretty safe assumptions in the case of controller input. But resetting breaks these assumptions: whenever you hit reset, that causes an immediate interrupt on most systems. So reset isn't just a 13th button: all the other buttons are passively monitored by the game, while reset isn't (and in the case of the SNES isn't even under software control). Now most TAS's that use reset just use it as a quick way to jump back to the beginning of the game: it doesn't really matter when reset is done, so the fact that it can only be triggered at the start of the frame doesn't hurt anything, besides perhaps the loss of 1 frame. But when you're talking about save corruption, the timing of reset matters a lot. TAS-ing, being based around the concept of frame advance, just isn't designed for this very well. IMO it's another reason why save corruption should be in its own category.