User Files from warmCabin

Upload All User Files

#638273167943803772 - Action 52 - Meong Checker

a52_meong_checker.lua
34 downloads
Uploaded 8/11/2023 2:13 AM by warmCabin (see all 30)
A simple script I used to validate my Meong submission. Action 52 managed to fill this basic RAM search task with a shocking amount of jank. You'd think I could just check if the Y coordinate decreases every 4 frames, right? WRONG. You go through 06 twice, for reasons unknowable. So my code has to account for that.

#638266238020257523 - Lollipop(s) improvements

lollipops improve.fm2
In 01:47.01 (6431 frames), 3807 rerecords
40 downloads
Uploaded 8/3/2023 1:43 AM by warmCabin (see all 30)
Some improvements to Lollipop(s). Interestingly, I only use the lollipop in level 1, now!
"You" is nymx, because you did some tests during the judging process. If you are not nymx, sorry.

Level 1

I straight up don't understand how the manipulation works at the beginning. Afterwards, though, I noticed that because you don't keep him on screen for as long, you get an extra candy wrapper enemy afterwards. By slowing down a bit, I was able to keep the pink guy on screen, spawn block that candy wrapper, load a little idiot who gets stuck in the wall, and then spawn block the enemy you jump over at 1464. Saves 2 frames.
Seems like you can't do spawn block strats with two enemies of the same type. It's like they reposition the same enemy to the new spot. Weird...
I found a new way to do the ending, which saves about 10 frames over yours.
50 frames saved over submission, 24 frames saved over your test.
No framerules between stages, so GG.

Level 2

It's actually a frame tie to smack that little pink think vs jump over him at 4029, but then an extra enemy spawns at the peak of the hill, slowing me down. Since I have to kill this newly revealed enemy, he can't be used to spawn block anything.
There's actually a frame perfect trick where you can kill the guy in the pit at 4219 without otherwise slowing down, but it just doesn't work out to climb the three groundpops at the end. You want to wait and jump across the whole pit.
I saved 16 frames at 4382 by waiting to hit the ledge instead of swinging my lollipop and waiting on the enemy. This was the only improvement I could find.
16 frames

Level 3

Couldn't find any improvements. Remember that you have to jump immediately because of weird collision on the starting platform, so I can't do the first pair of enemies optimally.
That second pair of enemies (that I creep up on) seems promising, but that only loses 8 green frames. A lollipop swing seems to lose about 15, and on top of that, more enemies will load and cause me more problems.

#638256808684726895 - Action 52 - full run proof of concept

full_run_test2.fm2
In 16:55.23 (61014 frames), 438 rerecords
System: Nintendo Entertainment System
3 comments, 46 downloads
Uploaded 7/23/2023 3:47 AM by warmCabin (see all 30)
I'd always pictured that an Action 52 run would consist of all 52 games played in a row, so here's a proof of concept for that. It consists of all the submissions and publications I could find, which is 8 by LogansGamingRoom and 2 by me. One of those is Meong, which I have not submitted yet, but I will. The submissions guide says if I submit more than 1 run in a day, they'll delete my account and set my house on fire, so I think I'm gonna wait until next weekend!
Ooze, Dam Busters, and Sombreros all had desyncs. I think they had to do with RNG, and I think the RNG is fed directly by your button presses (as opposed to, say, X subpixels or sprite slots), because I was able to despawn some enemies by wasting shots.
But what does this accomplish that a simple YouTube playlist wouldn't? Honestly, it just satisfies my misplaced sense of purism. We can't even get Cheetahmen on here because they played that on revision A.
Syncs on 2.3.0.

#638255087602508378 - Four Color a Dinosaur - Ultimate Ultimate ULTIMATE Submission Candidate

4color_2023_ultimate_ultimate_ultimate.fm2
In 08:09.93 (29444 frames), 17309 rerecords
70 downloads
Uploaded 7/21/2023 3:59 AM by warmCabin (see all 30)
Am I done yet? 62 frames saved over the previous submission candidate. These hot takes are backed up by FractalFusion's adjacency lists. Thank you for those!

Crested Duckbill (11)

Saved 24 frames by fixing a gray chest region that could obviously have been yellow. It was right along the way, too. Saved another 6 by swapping the gray and yellow mouth regions.
30 total

Mini Stego (13)

Saved 22 frames from changing that one spike to yellow.

T-Rex (14)

Saved 10 frames by changing that one toenail to pink.

#638253472486465559 - Four Color a Dinosaur Ultimate Submission Candidate

4color_2023_ultimate_ultimate.fm2
In 08:10.96 (29506 frames), 16683 rerecords
40 downloads
Uploaded 7/19/2023 7:07 AM by warmCabin (see all 30)
Can I be done with this yet? 178 frames faster than the previous submission candidate.

Spino (2)

Saved 24 frames from better routing around those fingers.

Tricera (5)

Realized one of those pupils could be purple, saves a measly 5 frames.
Implemented my routing improvements from morefix, saved another 9 frames. Still waiting for 10 frames on that final A press, though... This seems to be inevitable. There's just no way to fill all the grays in one pass without waiting.
Saved another 8 frames by doing the pink horn region thing first. All button presses are now frame perfect except for the very first (1 frame slow), the left pupil (2 frames slow) and the very last (10 frames slow). Cycling from gray back to pink also happens to cost 13 frames, so it MIGHT be possible to break even with a color cycling strat. Why is this game so good???
Just to clarify, you can't cycle your color during a fill animation. You gain access to A (fill), B (select color), select (select palette) and start (finish drawing) on the same frame. You CAN cycle the color while the pencil is moving, but that doesn't apply here.
22 total

Ptero (7)

Saved 15 frames using my hot take.

Mini Stego (13)

Non frame perfect delays:
 9, 3, 13, 5, 3
Looks like there are 33 frames of potential timesave from pencil routing.
First pass:
 6, 3, 13, 5, 0
Saved 6 frames by altering the toenail order. The 13 and 5 are gonna be tough to crack because it's the little sliver in the bottom right. Making that inner crotch region yellow is also really tough.
New route:
 3, 3, +13 for the loopback, +1 for the extra yellow transition
In this crazy new path, I switch colors early so I can have frame perfect transitions out of those tiny secluded regions.
Apparently adds up to 5 frames all told.
11 total

Proto (16)

Saved 106 frames using my hot take.

#638253299285074921 - Four Color a Dinosaur - new submission candidate

4color_ultimate_safe.fm2
In 08:13.92 (29684 frames), 15095 rerecords
47 downloads
Uploaded 7/19/2023 2:18 AM by warmCabin (see all 30)
Alright, here's the splice. Looks like it's 132 frames slower, but that's OK.
Differences from the published run:
  • Stego (1) - My fixes & improvements from April
  • Spino (2) - My fixes & improvements from April
  • Tricera (5) - Randomno's fixes from June (which are an improvement!)
  • Vinny (6) - My improvements from April
  • Ptero (7) - Unchanged, no hot takes. This has already been complicated enough for the judges...
  • Mini Stego (13) - Randomno's fixes from June
  • Proto (16) - Unchanged, no hot takes

#638252789553792720 - 4color_2023_morefix

4color_2023_morefix.fm2
In 08:08.95 (29385 frames), 15046 rerecords
42 downloads
Uploaded 7/18/2023 12:09 PM by warmCabin (see all 30)
My Four Color a Dinosaur publication has some pretty glaring mistakes, which people noticed after it got published, and Randomno is working to fix. Feel free to not believe me, but I did notice some of those mistakes back in April, and was working on a movie to address them. I noticed the toenail thing on #1 and the eye thing on #2, but #5 and #13 slipped by me a second time. Apparently I had some hot takes on #6, #7, and #16. Get out your Voronoi diagrams...
Compared to Randomno's submission, my #1 is 8 frames faster, and my #2 is 34 frames faster. I'll see if I can't work your fix into my improvements to #5; your #13 will be spliceable.

Here is my original notes draft file:

Stego (1)

I realized one of the toenails, which I considered to be at a corner with another, might actually be a proper adjacency. I went and recolored it ugly, and in the process, I saved 4 frames. I sacrificed 2 frames to show the eyes.
Future note: Apparently I saved another 4 frames by doing pink first.

Spino (2)

I actually bothered to redo this with the yellow spots routed into the main yellow pass. 16 frames. Then I realized I left both the eyes pink. How embarassing! I also realized I could color that nostril pink, saving some frames. This actually saved another 7 frames.

Tricera (5)

I managed to combine more long waits with long movements. 16 frames.

Vinny (6)

It's always vague with these large outlined regions, but I realized Vinny's right pupil could probably be considered touching his yellow body. I changed it to purple, which saved me 5 frames. Not only from avoiding the fill, but from avoiding a slight backtrack to reach it as well. I sacrificed 2 frames so you could actually see this.
Future note: I can't help but wonder why I decided to color it yellow in the first place...

Ptero (7)

I realized that little leg...socket...thing could be yellow, since it touches the body at a corner. 15 frames

Proto (16)

I changed my mind about that last splotch. It's only a corner, so I saved 106 frames for free.
Total = 169 (not incl. time/ent tradeoffs)

On Corners

Having seen all these drawings, you might think it's a little cheesey that the corners are allowed to be the same color. Well, that's just the way the theorem works. I believe the stego would be impossible if corners counted, but to really drive this point home, consider this dinosaur:
It's got 8 wedges all touching in the center, so you'd need eight colors. You could slice it into as many wedges as you want and there'd be no bound on the number of colors required. So yeah.

#637792162013061522 - Rockman 2 Basic Master - Snake 406 frame improvement

Rockman2BM-snake-406frame.fm2
In 03:23.00 (12227 frames), 11814 rerecords
Game: Mega Man 2 ( NES, see all files )
114 downloads
Uploaded 1/31/2022 8:56 AM by warmCabin (see all 30)
This movie file contains Shinryuu's improvements in Gemini, which I forgot to encode. He reduced lag and killed the second penguin without stopping.
Snakeman's stage is now done with a super fast new zip that doesn't involve the yoku blocks. We also added a couple extra Gemini scrolls to save even more time.

#637789489750540174 - Rockman 2 Basic Master - Gemini 176 frame improvement

Rockman2BM-cabin-176frame.fm2
In 01:53.00 (6778 frames), 26852 rerecords
Game: Mega Man 2 ( NES, see all files )
1 comment, 104 downloads
Uploaded 1/28/2022 6:42 AM by warmCabin (see all 30)
Here's how I saved time. See the YouTube description for more details.
  • Killing the penguins
  • Keeping to the ground. Walking is faster than jumping.
  • Jumping higher to fall faster past walls.
  • Careful lag reduction on the Galaxy Man dudes
  • Killing tadpole eggs to save a TON of lag
You can probably shoot the remaining two eggs and save a frame or two, unless that's really laggy. I forgot to try.

#57070174403385324 - Mega Man 2 Zipless - Cabin Route

cabin_route.fm3
In 18:47.64 (67770 frames), 64141 rerecords
Game: Mega Man 2 ( NES, see all files )
608 downloads
Uploaded 8/6/2019 3:42 AM by warmCabin (see all 30)
My route for Mega Man 2 zipless. It's the fastest of the three I considered, so it's the one I'm continuing with. It beats the RTA route by 104 frames, and aglasscage's route (Metal late) by 46 frames.
The route:
Q M B F H A W C
Comparison Video showing all three routes side by side
I made an improvement in Heat, which is contained in this movie. I didn't feel like implementing it twice and re-encoding both movies, so both this route and aglasscage's route are 8 frames faster than what you see in the videos.