Posts for Mizumaririn


Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
A few things regarding Japanese version. Damage boosting in US from fire/leaf to big costs 32 frames. Getting powerup from big to fire/leaf costs 32 frames. Damage boosting in JP from fire/leaf to small costs 48 frames. Getting mushroom costs 48 frames, then fire/leaf costs 32 frames. Therefore, each damage boost in JP costs 64 frames more than in US, plus the slowdown required. The P wing route: The glitched item in 5-1 can ~~only~~ be obtained using a P wing. After that is 5-F1. First, you have to get rid of the tail so there's 48+48+32=128 frames plus slowdown. There is no fire flower in 5-2 or 5-F1. 5 seconds lost as you cannot firekill Boom Boom. These are not accounted in Mitch's video. A quick glance, the fastest powerups to get is the 4th mushroom in 5-5 and the flower in 5-7. Though inability to firekill the red koopa troopa in the way forces a star be used in 5-7. As the platforms after are infavorable to charge pspeed. Star has to be used, costing 44 frames. In 6-F1, the fact that using a star is 1 frame short of getting through the spikes corridor, the P wing should be used. Otherwise, wait for the elevator and take damage once. IIRC it is half a second slower. But the longer damage animation should reduce the invincibility frames that have to land on the spikes much further. Or use a star and take damage. 6-F2 is ended with tail or small. The damage boost in 6-8 is not available as doing so will not get us the fire flower in 6-10. Meaning that we are playing 6-8 and the start of 6-10 as tail. In 6-8, you will have to run up the slopes, and having slighly lower than 56 throughout, and unable to get the 61 or 62 speed benefit from slopes. Or, damage boost in 6-8, do 6-9, stomp kill in 6-F3, get mushroom and stomp kill in 6-A, a jump around the 1st Piranha. This way, I guess it loses little time. Edit: Well Maru theorized a route keeping fire throughout W6 and is 1.9 seconds faster. So problem solved if this route is used. Into 7-4, now there is a trouble. As a P wing is used for 5-1, we run out of P wings. We may have to switch back to cloud 7-4. But hey, we can cloud 6-F2 to save the P wing for 7-4 and get the mushroom in 6-8 and having faster 6-8 but it still uses a cloud which DEFEATS THE WHOLE POINT OF GETTING A CLOUD IN 5-1!! There are limited P wings for TAS and trading a cloud with a P wing is not viable. Therefore, the leaf route: There are 2 leafs in 5-1, both at the beginning. P speed isn't available. Though in comparision to P wing with little ground it isn't much slower. In the end, I believe P speed can be charged before entering the pipe. After reaching the top, not having P speed is faster. Still, 5-1 as a whole would be slower. However, unable to firekill Boom Boom in 5-F1 and powering back up before 5-F2 still holds true. After 7-F1, you would be either tail or small. Small means you cannot clip 7-6 (although you can wall jump). Tail means you cannot clip 7-7 as there is not enough room to accelerate. That means tail has to be kept and 7-8 is played instead, which is slower than 7-7. 7-8 is either played without P speed, or damage boost, and then get the mushroom in 7-9 to do the clips. Both options costing a lot of extra time. Get the fire flower in 7-A and the rest is the same. So where to use the extra cloud? Probably to resolve one of the problems mentioned above. It wouldn't get the largest timesave possible, but makes a JP TAS with minimal P wings possible. The most suitable candidate is 6-F2, clearing which is 30 seconds. And it actually ends up the same with the P wing route. In RTA, 7-4 must be clouded. 2 out of 3 levels between 7-9, 7-F2 and 8-F are played. In TAS, cloud from 7-4 can be saved. 7-F2 and 8-F both are clouded. 7-9 is clipped with ease. 7-4 and 7-9 both become short levels. The next long level clouded isn't really long. So the benefit would be less than Mitch using 8-F for the calculation. Lastly, music box is not obtained from 5-1. To choose between fighting 5-HB and 7-plant, HB with the P wing should be fought. There is extra 10 seconds. The TAS route is different from the RTA route. Which items to use, in-level routes, etc. are different. While JP might be faster for RTA, it doesn't mean it is in TAS. Damage boosting being from leaf to small brings up a lot of problems more than the literal 64 frame animation cost. Extra P wings have to be consumed. All and all, the route would be reverted back to be similar to the second to last published movie [3028] NES Super Mario Bros. 3 "warpless" by Lord_Tom & Tompa in 46:20.30.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
I think "must be reproducable" will always be a strict rule. Or else no one can verify it. And why would you use not bizhawk? For wording, i would say just in the playground section, list the rules that do to apply to playground, then list the rules that replaced.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
I am concerned about would it be partly judged for leniency. Where it hasn't reached the exciting part, or long unentertaining setup just to setup something cool. Imagine smb3 without whistles. (ignoring DCPM workaround) you have to sit through 6 worlds without new content just to see cool stuff. Why should the other parts affect the viewing experience of the real playground - the main focus the author wants to show off? Unless the primary goal is to be fast.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
DrD2k9 wrote:
I'm not overly familiar with SMB tasing, but there's an apparent time loss at 2:00 into the encode. Mario bonks into the elevator platform in 4-2. Had you taken a slightly different route prior to that bonk, you wouldn't have been slowed by it. There is also a brief slowdown in 8-3 around 4:04 in the encode where you slow down to kill a hammer bros. I know there are frame rules involved with SMB, so i don't know if either of these slowdowns cause loss of a frame rule. It does look sloppy and suboptimal though.
Hi! The bonk in 4-2 actually sets up a trick called the wrong warp, and it is done by bonking or clipping into walls to push Mario to the right of the screen. Iirc the pace is accounted into 8-1 where it finally count the framerule. Idk is 3 bumps wrong warp optimal though. In 8-3, a slowdown is needed in order to avoid ending the level with 3 on the ones digit of the timer which causes fireworks and waste time.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
temp encode: Japanese version as base rom end has blank frames without input Link to video
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Welcome! I guess you are here inspired by some video explaining TAS. This is a TAS, where you don't do a sloppy run and call it done. Remember, you have total control of Mario. Avoid bumping into walls if you can. Check out what tricks saves time. At least nice job executing the flagpole glitch. You can check out many more glitches from Wiki: GameResources/NES/SuperMarioBros. You can also learn from the published movie on the site [1715] NES Super Mario Bros. "warps" by HappyLee in 04:57.31. Also note that TAS and RTA may have common tricks, but different ways to execute.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
7 days 1. Time for viewers to give feedback 2. A judge should not be urged to judge. Require judging instantly can make judges overwhelmed by the work. Giving many days to judges can give them a clear mind.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Bad game choice
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
It is. For convenience, a savestate can be provided. Otherwise, a movie file used to create that savestate can be provided, just like how movie file providing SRAM work. Can this tier be "branch extentions"? I think, the difference is, the encode starts at the middle instead of power on.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Sorry. Thanks. But which pages link to this page? Only in special categories in Movies page... still thank you for redirecting me.
pronouns: Mizu/Mizu
Post subject: The Extremes of TASVideos
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Page idea: The Extremes of TASVideos Lists the shortest movie, the longest movie, the game with most movies, the user with most movies, the most viewed movie, the movie with highest rating, etc.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
McBobX wrote:
Can you do it in 1 frame? Or or, can you just show credits?
No. First of all, there is nothing you can do to cut the first 10 frames. The CPU is busy clearing RAM. Non-maskable interrupt is disabled and controllers aren't read. Second, it couldn't be done in 1 or 2 controller-reading frames due to limiting options what you can do on each frame. Here is an explanation by 100th Coin.
100th Coin wrote:
I think I can prove that you can't make a faster TAS of this game. The big three requirements are setting the stack pointer, setting the NMI mode, and jumping to $B85A. The first frame of input writes the JSR to $9000, the next two frames fix the stack pointer and take that jump. Let's ignore the first ten frames, as there is genuinely nothing you can do on them. The controllers simply aren't read. Frame 1 is very rigid, as it requires stalling for 100 polls. You could change what you press for your final inputs on controller 1 and 2, and nothing more. These bytes don't get read as code, but they do get stored at $0000 and $0002. Frame 2, you have no reason to stall with subframe poll mashing, as that would only let you execute less instructions. That means you only have 2^14 options here. Initially I was going to say that you are limited to controller 1, as a single frame can only get you to about address $DB. I did have an idea to get around that though... You could write a branch on this frame that would let you skip enough bytes so that you can execute $F5 through $F9. EXCEPT You would end up writing another branch at $F5, as you can't change the inputs on the same frame, making it impossible to write anything useful here. Therefore, you really are only limited to controller 1 on frame 2, and you're only able to write instructions with no d-pad conflicts, and the opcode must match the operand. (because this is the first frame that properly records inputs. According to the game, the newly pressed inputs equal the total held inputs. Opcode must match the operand) There is nothing that could be written on frame 2 to win the game.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
NEVERDOORS wrote:
you may not like it, but this is what the ideal speedrun looks like. gg!
Yes. Unfortunately being the fastest is not necessarily entertaining. The same goes for RTA speedrunning, where for some people, it is actually not fun to grind for world records. Although entertainment value is important on this site, this definitely demonstrates superhuman skills. There is no way a game completed under 13/60 seconds not documented on this site!
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Movie rules:
A speed-oriented movie must beat all existing records If your movie is going to beat something, be sure it beats it. If your tool-assisted movie is slower than the non-tool-assisted world record for the same game, aiming for the same goals, your movie will be rejected.
Movies not intended to be submited should go to the user files instead.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
feos wrote:
This is the same definition of 100% as on SRC, which is actually "all levels", right?
Yes.
feos wrote:
Hacks specifically get handled like Moons branches, which means they can obsolete one another. For this one, it looks like a better (and more popular) SMB3 Kaizo hack has already been published and got much better feedback. So improving that movie instead would indeed be a great idea.
In fact, I have moved on to one of the most popular hacks. I don't want to ruin my surprise. vvvvvvvvvvvvvvvvvvvvvvvv
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
BigBoct wrote:
Around 1:35, is it possible to use the pipe glitch to enter the pipe without waiting for the Para-Beetle?
Pipe glitch can only performed at the right side of the pipe to my knowledge and I can't reach there earlier.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Well, I just wanted to TAS a hard game and I found this hack. I never considered the aesthetics. I feel successful finishing this, and I'm here to share my joy.
pronouns: Mizu/Mizu
Mizumaririn
Other
Experienced Forum User, Published Author, Player (238)
Joined: 2/26/2020
Posts: 43
Location: Super Bell Hill
Considering this is a small hack and low popularity, should I cancel the submission instead of being rejected?
pronouns: Mizu/Mizu