Posts for warmCabin


1 2
7 8 9 10
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video Thanks to Shinryuu's advice, I was able to make some improvements to my run. I've also completed up to Heat (just gotta finish up the fight). Userfile: User movie #50491715316358744 Discussion thread: http://tasvideos.org/forum/viewtopic.php?p=475848#475848
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
User movie #50491715316358744 Link to video I've TASed up through most of Heat. Just gotta finish up the boss fight. Some stuff in Air, Quick, and Metal has changed, so I re-encoded the whole thing.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Ah, yup. I'm doing zipless, but not glitchless. In realtime runs of this category, I regularly execute ladder popups, bird egg despawns (is that even a glitch?), and the Boobeam quick kill glitch. Zipless is more of a "No Major Glitches" type of thing, where we ban only zipping, screen wrapping, and delay scrolling. I do like your "no faster than you could normally move" rule, though. I think it applies here.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
I posted this on the Mega Man Discord, but I wanted to confirm it here with you folks. This is where I'll be submitting my run, after all! So, let's talk definitions. What exactly constitutes a zip? Clearly this does: But what about this? You can't see it, but here, Mega Man pops up from 94.08 to 94.00--3% of a pixel. Seems like pretty standard behavior for the collision code, if you ask me. But before you answer, look at this: A lot of jumps in this run end up looking like this. You clip inside the block by less than half a pixel, then get pushed out on the next frame. Should this be considered a zip? Redoing large portions of my run to be slower just so I can avoid this particular edge case seems absurd. Especially when you consider realtime runs; how can a judge tell if I popped up 3% of a pixel on a particular jump, or if I landed "normally?" My proposed definition is this: if the floor/wall ejects you less than one whole pixel, the collision code is behaving as normal, and you are not zipping. This definition allows me to get good subpixel precision in my jumps, while banning the intended exploit: jamming your body into the ceiling and blasting off at inane speeds. It also provides reasonable criteria for RTA judges to look for in any ambiguous runs. Regardless of whatever consensus is reached here, I've decided to jump for an extra frame in the particular scene that I showed, to circumvent the issue. I'm waiting for that Item 2 to spin up, so it doesn't really matter. I just wanted to clear this up before moving on.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video We already have a pacifist TAS of this game in the Vault, and an unpublished one from a decade ago. This one imposes the additional restrictions of defeating as many enemies as possible on hard mode. What do you think? Different enough to be worth continuing? Userfile: User movie #49606539655244104 Discussion thread: Thread #10903: Silver Surfer
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
S4FHFB%, anyone? Link to video User movie #49606539655244104 I know we already have the pacifist one and the unpublished one from a decade ago, but perhaps these new restrictions (hard mode and max kills) impose an interesting enough challenge to make this category worth a watch. What do you think?
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Aha! I knew something was amiss! It turns out switching weapons during sprite flicker (not lag) can cause your teleport animation to last 10 frames longer. With a better menuing frame, the TAS inputs save 2 frames to mine--3 if I can manipulate away a pesky lag frame. I feel much better now. I seem to recall saving one frame to them in a different movie file, so I want to look into that just a bit more before I move on. It was probably a similar mistake.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Fantastic run! World 5-Tower must have been MISERABLE for you. Did you have some kind of tool to help you see where you were going? Yes vote from me...even if your new improvement is going to be outdate this run like a day.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Here I go posting in this thread again. Link to video The Item 2 section can be improved by a couple frames, it looks like. Userfile: http://tasvideos.org/userfiles/info/49364838704578513 Discussion thread: http://tasvideos.org/forum/viewtopic.php?p=473913#473913
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video I finally decided to move on to Metal Man! This stage saves 50 frames over v1.0. Regarding the Item 2 section: unless I've set something up wrong, I believe I've saved 9 frames over inputs copy/pasted from the 2010 TAS. It's because they couldn't quite catch up to the tip of the Item 2 after jumping over the tin can man, whereas I could. That's the only reason I can come up with for why that happens!
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video Yet more timesave. I shaved 33 frames off the previous buster fight with an alternate strategy. Every time I think I'm ready to move on, I come up with another idea that saves yet more time...so I think I'm gonna sit on this one for a couple days before I call it optimal. I plan to post a userfile with all my attempted strategies in the branches.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video Now this is a cool fight! There's only one problem: it's 95 frames slower than the buster fight... There's no way to deal damage to him when he's charging, so you're forced to wait until he hops. That's the main reason this fight is so slow. Speaking of buster fights... Link to video I've managed to save 18 frames from the previous fight, which saved 18 frames from the previous fight. Weird coincidence! I can't seem to find a way to get behind him without taking damage or wasting a lot of time, so this seems to be the best solution. Letting him charge to the left (like the 2010 TAS) isn't an option, because buster shots have much smaller hitboxes than crash bombs, so I can't shoot him when he's down there. I also had trouble dodging the second boomerang shot in a timely manner; perhaps I could manipulate a higher jump and see if that works out.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
No worries! And you're right, he takes 14 hits from either weapon. The air fight would look something like this: Link to video I wasted 4 shots throughout the stage for entertainment purposes, so in reality, I have more than enough ammo to pull this off. Dodging those quickarangs presents a problem, since it allows Quick Man to get up on the ledge, which is bad. Once again I manipulated a high jump so I could dodge the first quickarang throw. The 2010 TAS does not, so I'll have to figure out if that's doable in this run. Not sure if they use the boss timer pause glitch thing to manipulat that jump.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
As far as the snipers go, it's extremely close. The air shooter route saved 6 frames, until I discovered an optimization that only works for the buster route; now buster is 1 frame faster. You can check branch 8 in the userfile if you want to see for yourself. I plan to make a comparison video at some point, though. As for the boss fight...is that feasible? Can you defeat Quickman in a reasonable amount of time with the air shooter? The only comparable TAS I could find was the 2003 TAS, where they switch to buster for the fight. I think the way to go is to shoot him at the beginning of a high jump, so the second tornado hits him on his way up, but I haven't fully explored this idea.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video Decided to start from scratch and fix my mistakes. I've already saved 142 frames! This movie plays through Air and Quick Man. I think the Quick fight is improvable. http://tasvideos.org/forum/viewtopic.php?p=473484#473484
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
v2.0 is in the works! I felt motivated to start on this rather than finish v1.0, although I still do plan on finishing both versions. v2.0 is the version I ultimately intend to submit. I've already managed to save 142 frames in Air and Quick alone. Currently, I'm not super confident about my Quick fight, even if it's 18 frames faster than v1.0. I don't know if the pattern used in the 2010 TAS will be viable, because buster shots have smaller hitboxes than crash bombs. userfile link Link to video
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
I believe I've discovered a new subpixel optimization technique! I don't see it mentioned anywhere in this thread, and it definitely isn't used in the 2010 TAS, but it's possible the supixel savants in here have discovered it since the previous TAS. I apologize if this is already known. User movie #48877130464294925 The details are in my userfile, but the gist is this: it's not always optimal to brush up against a wall when turning around. My movie demonstrates this by shaving 3 frames off the 2010 Flash Man segment (technically it's 1 frame + a door frame rule); I'd bet there are about 10 frames to be saved with this throughout the run.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Just to post my thoughts: I'm looking into handling Wily 2 more like the regular TAS. That is, Wood for the drills, Time Stopper for the mashy spike plates. If this is faster, it will leave me with a total of 2 extra Item 1's. I plan to use one in the alien fight to get onto that secret ledge. I might use the second for a quicker ascent in Boobeams, or in the infamous "hold up" screen in Wily 1. Or, maybe I'll find a way to get good MadWidth/MadHeight on that one unavoidable bullet in Wily 4, so I can use an Item 1 to dodge the final Sniper Joe like I originally intended!
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Link to video Stopped in the middle of the Boobeam's, because I'm a jerk. Considerations for the Boobeam fight and the Wily 2 drills are forcing me to reevaluate my Item 1 usage; do comment if you have any ideas. http://tasvideos.org/forum/viewtopic.php?p=472235#472235
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Progress update, and a couple questions: Link to video I'm currently deliberating over the Item 1 route, here. Do you think it will be worth it to take the extra hit from the Sniper Joe and use the spare Item 1 in the alien fight? That seems like the best Item 1 usage to give up, if I had to pick one. I was also wondering about the route in the boss rush. Why is it that real-time runners do an order that minimizes weapon switches, while the TAS does an order that minimizes distance? Shouldn't one route theoretically be faster in both cases? Edit: Shit. The Guts Tank fight...I just realized you can shoot him every other frame. Shooting from a distance, I genuinely believed he could only be damaged every 3 frames. Each quick shot you see represents the earliest possible frame he can be hit from that position. Interestingly, I realized this mistake when I noticed that, in the current "zipful" TAS, the Guts fight sounds different. You can hear one more frame of audio for each shot!
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
I found a good frame to get hit on the upper ledge, but it forces me to delay my crash bomb by quite some time. I also found a good frame to get hit by the final Sniper Joe outside (leaving me with an extra Item 1!!!!!), which ends up being 338 frames faster... I imagine these arbitrary frames are going to be different in the final run, so I may decide to just pick one and move on for now.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Progress update! Perhaps unsurprisingly, I'm stuck on the Boobeams. finalfighter's script is proving to be extremely useful, but I just don't know how to manipulate these "Mad" values to be what I need without wasting substantial amounts of time. They seem to cycle through these weird outlier values that I don't know how to make them hit. I'd appreciate any pointers you MadWidth Masters might have.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
So I was pushing on through the Wily stages, and I have a question: Look at the current TAS (exactly 20 minutes in, if the link didn't work). The way they coordinate warp there--should that be considered a "zip?" I know that same trick in Cut Man's stage is banned for Mega Man 1 zipless, so perhaps I should try to find a way around it. But I'm not really sure. Edit: Hmmm...I may have answered my own question. Any% RTA runners go for this coordinate wrap, but Zipless runners do not. I'll assume it's banned in this category.
Experienced Forum User, Published Author, Player (248)
Joined: 3/15/2018
Posts: 229
Location: United States
Proper aspect ratio + more pixels than I know what to do with = a much better encode. Thanks!
1 2
7 8 9 10