Advertisement
Baronhaynes

Mega Man 5 Theory TAS #2 notes

Mar 19th, 2015
308
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 17.82 KB | None | 0 0
  1. Rockman 5 Theory TAS by BaronHaynes in 33:53
  2. Done on FCEUX 2.2.2
  3.  
  4. Goal:
  5. Create a perfect playthrough video using only current or theorized RTA strats for Rockman 5.
  6.  
  7. Parameters:
  8. 1. Stick to most likely stage RNG to show the optimal movement in an average playthrough (no Gyro or Charge ammo drops).
  9. 2. Perfect or near-perfect boss RNG to the best that I could manipulate it.
  10. 3. Movement should resemble what's realistically possible for human players. No 30Hz mashing, micro-slidejumps, frame-perfect turnarounds. Include only what I or other players were capable of doing RTA with some degree of consistency.
  11. 4. No arrow or coil zips for now as no one is attempting them.
  12.  
  13. How did I do:
  14. Pretty well. I think I pushed the limit on "realistically possible slides" in a few places, such as high ledge micro-slidejumps (I tried to impose a 5-6 frame minimum on any slidejump). Arrow startups are tight but realistic, with a similar limit of a 6 frame jump to land on the arrow again. I think virtually all of this movement and these strats are doable by the top players of the game, or that someone could realistically try in the future.
  15.  
  16. All boss fights are based on current or theorized RTA methods and have frame-perfect hits, which results in some faster fights than anyone has seen in real time. I didn't cut any cycles earlier than I believe a human player is capable of, though whether we'll ever see this Napalm fight in a run is anyone's guess.
  17.  
  18. Some of the strats I chose to include aren't ones I'd recommend trying (or plan to do myself), either due to difficulty, consistency, or risk/reward concerns, but I think they're all doable if a runner put the work into them. I've included notes on rooms that may need further explanation or seem questionable.
  19.  
  20.  
  21. STAR
  22. ----
  23. Asteroids - It's hard to do accurate lag measurements in this room or to know how charging affects it. The ideal is probably to start charging midway into the room.
  24.  
  25. Second Room - The instajump over the cannon after killing the second thumper is a 1 or 2-frame trick depending on your fall speed and how close to the ledge you are. It can be modified to a near-instajump and still save frames. Cut it as close as you like.
  26.  
  27. Checkpoint Room - Instajumping over the cannon here depends on a perfect 1-frame jump over the met with enough room to land. Not recommended, but possible.
  28.  
  29. Met Room - The third partial slide before the second spike jump is also insanely precise.
  30.  
  31. Torpedo Boost - Dboosting the torpedo is 15 frames slower than jumping over it unless you get fast gravity and optimal movement, where you can save 10-12 frames.
  32.  
  33. Boss - Star Man won't fire his shield if you are more than half the screen length from him when he starts a jump from the ground. That's the only surefire manipulation I know of. Star Crash's hitbox is much smaller than it looks - only a few pixels larger than the boss's and no part of it touches the star sprites.
  34.  
  35. Gravskip - I've heard conflicting reports, but I get it more often when I mash A compared to pressing nothing at all.
  36.  
  37.  
  38. GRAVITY
  39. -------
  40. Screen 1 - Killing both eye guys with buster and switching to arrow has very tight timing to avoid the gumball machine, but it allows for an extra slide in this room.
  41.  
  42. Screen 2 - The setup in the previous screen means no charged buster, so the first eye guy has to be killed with two arrows. A+B jump into the gravity flipper to hit him, then turn and place the arrow to the right. This is a 4 or 5 frame trick. This whole strategy saves 18 frames (and is pretty easy to do).
  43.  
  44. Lupin_wal Boost - From what I can tell, this is a 2-frame trick, hitting the red spine at the absolute peak of your jump to flip gravity. With practice, it's very doable and saves about 15-20 frames.
  45.  
  46. 2-Tile Ledge Slides - I was a bit overzealous in chaining a few of these together, but if you have a chance and feel up to it, even clipping the corner can save a few frames over walking.
  47.  
  48. GYRO
  49. ----
  50. Buster vs. Arrow - They are virtually identical, but arrow is harder to do optimally. My comparison video includes some optimizations that no Japanese players currently do, making it slightly slower than the buster video.
  51. Comparisons: https://www.youtube.com/watch?v=v0gl58C2iXg
  52.  
  53. Large Ammo Drop - Using the buster route and switching to arrow for the last half of the stage with the refill saves about 30-40 frames.
  54. Comparisons: https://www.youtube.com/watch?v=pBGXqU9jIkc
  55.  
  56. Letter Room - You have to very slightly delay scrolling the second dropper onto the screen after the first in order to get between them, and movement must be insanely good. It saves 10-12 frames but it takes a while to get the hang of it, and to be able to see whether or not you're going to get it and abandon the strat if not.
  57.  
  58. Satellite Room - Killing the fifth satellite with a high jump from the second platform is faster than doing a partial slide jump to kill it from the third, as it optimizes slide distance on this screen.
  59.  
  60. CRYSTAL
  61. -------
  62. Dropper Room - Using a charged buster to kill the bouncing head and firing an arrow from a jump saves 10 frames in this screen and offers a chance at having 7 ammo for an extra arrow up the ladder on the next screen (another potential 15 frames).
  63.  
  64. Shield Attacker Room - This Rush Jet slide is theoretical, I've only been able to do it once RTA.
  65.  
  66. E-Tank Room - Getting the E-Tank costs 20 frames over jumping past it on the right side. Going across the top is 18 frames slower.
  67.  
  68. M-Tank Room - Shooting the lag bullet seems to be even with not shooting it. I tested this several times and never saved any frames.
  69.  
  70. NAPALM
  71. ------
  72. Met Cannons - Optimally killing the second met cannon and sliding immediately as he dies saves 14 frames over an optimal damage boost (slide into him, then through him).
  73.  
  74. Cave Arrow - Both arrow setups seem to come out the same in terms of framecount, so just pick your favorite. Note that this screen is notorious for triggering the arrow glitch (where you become unable to bump off the front of the arrow in Napalm and Charge), so adjust your setup to avoid lag in certain spots if it's happening regularly.
  75.  
  76. Letter Room - There's a 2 or 3 pixel window where you can safely drop into the alcove without falling past it. The arrow has to clip through the missile launcher's head or just barely graze past it. There may be subpixels involved with this, but I get it more often if I fall right, then left into the alcove. It saves 12 frames over a perfect ledge clip, and more over landing further left.
  77.  
  78. Missile Room - The last slide jump into the screen transition only saves 2 frames done perfectly, and loses time to lag otherwise. I just slide here.
  79. Spike Room - Same for the middle slide jump onto the ledge here.
  80.  
  81. Boss - I have no idea how to do a fight this clean in real time, but it is theoretically possible with almost frame perfect hits.
  82.  
  83. STONE
  84. -----
  85. Stone Mets - Killing the mets is risky due to the babies' jump height being RNG. You can get most patterns by shooting fast and close while they're still clustered. I did RTA tests of this screen for an hour and found that the first met saves about 11 frames optimally and the second only saves 2-5 (and seems harder to kill). Only really clean kills seem to save time over optimal jumps.
  86.  
  87. Screen 2 - I compared all four ways to do this screen: killing the met, sliding into and through it, and both 1- and 2-jump killing the bat with a charged buster. Killing the met can be 10 frames faster, but in RTA tests everything I tried was nearly the same. You can just slide into the met if this screen is giving you trouble and lose almost no time (but you have a higher health drop risk).
  88.  
  89. Screen 4 - Extra mini slides are possible here but only save a couple frames if you're insanely good at them.
  90.  
  91. Stone Arrow - This is a very precise trick that only saves 18 frames all added up. The only reason it's in here is because it can be done consistently with practice. Getting past the rock thrower is precise but not exceptionally hard. The hard part is getting to the end of the mouse room and up the ladder with 7 ammo remaining. Place the arrow using the seam in the background tile as a cue and do a full jump off it immediately to ensure you save as much energy as possible. If you cut any part of this trick, you lose all time advantage.
  92.  
  93. Not Stone Arrow - If you kill the rock thrower and arrow up the ladder like usual, you can jump off another arrow up the ladder at the end of the mouse screen. This saves 10 frames over using the ledge to jump up the ladder (and you should comfortably have ammo for it).
  94. Comparisons: https://www.youtube.com/watch?v=JaGIlj7hTcE
  95.  
  96. 3 Right Arrows - You have to tink off (or kill) the shield guy with arrow 1 and kill the middle telly with arrow 2 to save time doing this. Walking to the end of the arrow on this screen loses frames to lag, even when accounting for fewer enemies on the screen. I only saved about 10 frames doing this, so I'm probably not doing it in runs.
  97.  
  98. Rush Jet Ledge Slide - I accidentally left this in instead of the more optimal strat of shooting the third telly from the spinning platform (10 frames difference). The lag from leaving this enemy alive eliminates the advantage of sliding from Rush to the ledge.
  99. See comparison here: https://www.youtube.com/watch?v=0xtVjya9Bx0
  100.  
  101. Boss - This isn't the most optimal possible fight, but manipulating the God Stone was a nightmare.
  102.  
  103. CHARGE
  104. ------
  105.  
  106. First Train Arrow - The lowest possible pixel is the ideal placement, with one slide and no edging forward. Edging forward at all on this screen seems to lose time to lag. It's better to be on the left side of the arrow even with the no-slide placement and let it carry you to the opening without moving.
  107.  
  108. Chicken Room - Killing the first chicken in one jump is not as hard as it looks. I use a visual cue from the background and try to fire at its midsection on the way up to ensure 3 hits in time.
  109.  
  110. Checking for Drops - Killing the first mouse with an early jump only costs 5 frames. Checking the last mouse by firing on the way up to the letter costs no extra time, but you have to be careful to not hit the lower mouse or move left when avoiding it (thus scrolling the drop offscreen and respawning the last mouse). There may be other enemies worth checking.
  111.  
  112. Drop Comparisons:
  113. https://www.youtube.com/watch?v=MR3BgxAhLK4
  114.  
  115. Crystal vs. Buster - Switching to Buster is faster than using Crystal, even with optimal ammo drops. They're about even in execution, but Crystal requires a pause to switch to Stone, which adds about a second.
  116.  
  117. Boss - The triple shot is very tight, but with this setup, you can go for it and lose no time against the old fight if you don't get it. This fight also eliminates the damage boost during the refight, saving 20 frames.
  118.  
  119. WAVE
  120. ----
  121.  
  122. Spines - Jumping the last spine rather than using an arrow saves 11 frames due to lag.
  123.  
  124. Double Bubble - Depending on your setup, getting the arrow to carry you can leave you with enough ammo to do 3 arrows. This is definitely riskier, but you can do it even if you aren't all the way to the left on the small bubble. I'm not sure if I want to do it myself yet, though.
  125.  
  126. Boat Skip - Two arrows save time over three due to lag as well, with an optimal jump into the screen transition. It's actually possible to space two arrows so that you can slide into the screen transition without jumping, but I've only done it once RTA and it was by accident, so I included a near-optimal jump here instead.
  127.  
  128. Autoscroller - Kill the dolphins in front of the letter to reduce lag when the jetski Joe comes onscreen.
  129.  
  130. Octodad - This is the fastest version of the fastest fight I could do myself. If you could do 14 damage in the first phase, you can actually kill him before his second shot, but the most I could do was 13 (and that was rare). I tend to do 12 and 8 when I can get the 2-cycle.
  131.  
  132. Boss - Wave Man's water spouts are spaced so that you'll run into them if you walk straight or slide. The location is perfectly chosen based on your position when the attack is cued up. There's not much you can actually do with this in a speedrun, but the locations aren't randomly chosen somewhere in the room.
  133.  
  134. DARK 1
  135. ------
  136.  
  137. Screen 1 - The jump over the low satellite is a 2-frame window. Not recommended, but certainly possible. It only saves about 15 frames.
  138. - This method of killing 2 satellites with lemons on the way down reduces a lot of lag and saves ~20 frames over charging early and jumping under them. It's also much easier.
  139.  
  140. Missile Droppers - You can get 4 slides and then a neutral jump over the first dropper, which cuts out a slide jump and saves around 12-15 frames. This isn't too hard with practice.
  141.  
  142. Boss - I'm not sure if I optimized this fight as well as I could, but I had to wait for his shot to leave the screen before firing a wave to preserve the AI loop. Cancelling all 3 wave sprites on the final hit saves around 40 frames.
  143.  
  144. DARK 2
  145. ------
  146.  
  147. Double Gap Jump - You're as likely to lose time to this as gain it, unless you're really good.
  148.  
  149. Train Room - Jumping between the first met train's bullets but under the flying enemy is stupidly precise.
  150.  
  151. Conveyor Room - Killing the two skull droppers at the end of the screen seemed to create more lag in my tests. I never had a pass of this room where it saved time.
  152.  
  153. Skull Joes - You can actually go too fast in this room, getting to the second joe before his crystal shot forms. This TAS has to wait to execute this strat correctly.
  154.  
  155. Gumball Room - Ohon's method of killing the two satellites with lemons is about 15 frames faster than using a fully charged shot, due to lag.
  156.  
  157. 4 Thumper Room - I've discovered a better RTA method for getting past the second set of thumpers - kill thumper #3, but don't press B again until shortly before you slide. Thumper #4 lives, but you get under him and over his bullet without having to wait. The lag this causes is negligable.
  158.  
  159. Tiger Room - The alternate first tiger kill is to do a full jump off of the ladder (but not immediately, or the flying enemy's arc will lead it into you), release your charged shot at the height of your jump without pressing B again, and do 2 non-consecutive short platform slides in the same manner as in Stone Man. This is easier than trying to get 2 consecutive ones.
  160.  
  161. Boss - By using a damage boost and walking through the boss during i-frames, you can optimize beat hits while still killing him in the absolute corner. This saves around 24 frames, but took me a long time to get consistently RTA.
  162.  
  163. DARK 3
  164. ------
  165.  
  166. Spinning Platform Rooms - Room 1 requires a delay or suboptimal movement to get past the telly. Killing it didn't solve this, since Mega Man will be stuck in a shooting animation on the ladder. Getting through room 2 without killing anything saves a handful of frames.
  167.  
  168. Met Cannons - I did a lot of testing in this room. In a TAS setting, this method is only about 10 frames faster than a perfectly optimized version of the bjw/Ohon strat (including all pillar slides, fast arrow startup, and an additional ladder arrow). But in RTA comparisons it tended to be about 20-30 frames faster when I did both optimally. It has a higher chance of death due to the second cannon kill and getting ammo drops from the third one.
  169.  
  170. Checkpoint Arrow - bjw's arrow is fastest here, but there's a risk of placing the first arrow too low to clear the spikes.
  171.  
  172. DARK 4
  173. ------
  174.  
  175. Tower - I didn't include shooting the 4th block in the first cycle because I didn't think anyone was doing it, then found out Ohon did. It's an option.
  176. - The game will drop any input on the frame any block is destroyed. It has nothing to do with the actual falling tower animation.
  177.  
  178. Boss - I shoot to manipulate Dark 4's RNG and keep him from jumping. This isn't the best possible fight, but it's the best I could get with my knowledge of RNG manip at the time.
  179.  
  180. WILY 1
  181. ------
  182.  
  183. Spine Room - Scrolling the screen to the right saves 8 frames in a TAS setting, but it's more difficult RTA and my comparisons there were about identical.
  184.  
  185. 5 Pukapelly - Jumping off this ladder is pointless as you have to walk before sliding to get under these guys once they move (yes, apparently that's their name).
  186.  
  187. WILY 2
  188. ------
  189.  
  190. Water Wheels - Sliding all the way up to the walls is likely to be faster RTA than jumping early or trying to slide-cancel.
  191.  
  192. Darkman Jump - I found this much easier once I started changing my grip during the fall (index+middle on B and A).
  193.  
  194. Stomper 1 - You need to move Rush Jet up for about 4 frames total to make this jump. You don't want to move upward on Rush more than you have to, since it causes lag.
  195.  
  196. Boss - I didn't get an optimal Q9 fight, but I took it since I got the early 2-hit.
  197.  
  198. WILY 3
  199. ------
  200.  
  201. Gyro - You lose 25 frames if he dies in the air.
  202.  
  203. Stone - The weird sliding I'm doing is to manipulate an early jump, which saves 50 frames. Shoutouts to someday getting double God Stone in a run.
  204.  
  205. Wave - Damageless hits on Wave are frame-perfect, but I included this one because I've gotten it several times RTA.
  206.  
  207. Wily Press - Manipulating all 3s on Press is nearly impossible as far as I could tell, as there's no way to generate lag during the fight and no way to change his RNG without pausing. I took 7773, the ideal pattern if you get no 3s during the fight. All 3s would be 1.2 seconds faster than this.
  208. -A full breakdown of Press comparisons is here: http://pastebin.com/TeEHKv3k
  209.  
  210. WILY 4
  211. ------
  212.  
  213. Machine - I feel like this fight still hasn't been truly mastered. I first did a version where I leapt into him during an initial vacuum phase and got perfect point-blank hits while dboosting, but this didn't seem to reflect the current RTA trend, so I went back and did this version instead. I think opening the fight with a jump toward him (to collide if there's a vacuum) allows you to adapt to any pattern, but it needs more research.
  214.  
  215. Capsule - 6-cycle depends on starting with an early hit, getting 2 hits and a track off of a blind spawn and not losing any energy, then getting a low spawn for the last cycle.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement