Posts for MUGG


Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
I've been meaning to ask how wall collision checks work in this game. Is it that Link has multiple points inside him that aren't allowed through walls? Are all except for one point disabled while a chest is opened?
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
The finder responded now saying where in the level he died. http://i.imgur.com/mt6S1oj.png He said he played on 1.1.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
I have tried quite a few things now to try to get a butterfly clip to work, but nothing seems to work... I'm going to give up on it and TAS the level like in the previous run. I found that a clip works if both butterflies push you from below. This will depend on whether or not I can manipulate their movement (it depends on what you press, your position, and your form from my observations. There is probably some strange algorithm that isn't known yet). --> Here is a pic I made to show the potential of clipping with butterflies:
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
It's been a good week and the finder still didn't provide any useful information. There is not really a way to investigate a bug with such little information, so I'm not going to look into it further. It may as well be a fake report (because there are fakers - remember the Earthbound Onett clip fake?). It's unfortunate because I was kind of looking forward to breaking SML1.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
I'm thinking if I should host a competition like that.
I think I will not do it if there is a lack of interest
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
I think it would be more impressive if No Doors can be done in a category that accomplishes more of the game. How would max% no doors look?
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Lv4 done Yup, rerecord count has doubled due to extensive clip testing and it was worth it. I like how the goku clip worked out in the end. Prepare for butterfly clip in Level 5 (jk, but who knows). 1206 frames saved, not taking into account the emulation difference. When taking emulation difference into account, I think it's about 1422 frames saved. Run is going to be 5:1x pace now. ^_^ @BioSpark, thanks
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
dunnius wrote:
For technical, how am I supposed know how much work went into the TAS? It is very hard to tell, even from reading submission comments. It doesn't make sense. And it doesn't make it clear how to rate, which leads to inconsistent ratings.
Which is why I took a little time to rate my own movies, for fun, because I do know how much work went into my own TASes. But I think some people already agreed (on IRC or here) that the rating system could need a rework to make for more consistent ratings. This should probably be discussed here.
Gay wrote:
So what MUGG is saying seems every movie that is well optimized deserves a rating like 9/10. And since each film must be well optimized, does not that mean that every technical evaluation would be really high?
Yes, pretty much. But everyone seems to have their own definitions of what an "8", "9" or "10" is. For me, 8 = kind of improvable 9 = hardly improvable 10 = perfect I rated my own movies in the first post to defy some of the bad ratings I received - but I did it mostly for fun. Again, as said, I think we should use this topic to discuss a rework for the rating system, or maybe we can come up with a clearer definition of what an 8, 9 or 10 should be. For the record, I did not even take into account the "techniques" used in my movies. I only considered the frame-perfectness of the runs.
Post subject: Movie ratings
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
This is about the movie ratings on TASVideos. I - and some others - have uttered that they feel that (their) published TASes are unjustly rated. Everyone comes up with a different meaning of how good a movie must be perceived to receive a certain rating. Maybe this aspect could be discussed here. But I have also been meaning to publicly give a technical rating for my own currently published movies and justify those ratings. Since this is going to be just my own TASes, I felt that this topic didn't have enough merit to be posted in the General forum section. Please note that I don't want to call names of people who I think rated unjustly. Please don't start flamewars. --- Kirby & The Amazing Mirror 100% in 58:10, received 8.4 technical rating. I would give it a 7.5, because of known possible improvements (soft reseting, better boss strategies, some route changes, optimization in general) that can improve this run by about 3 minutes. Bubble Ghost in 4:41, received 7.1 technical rating. I would give it a 8.5 because this has been rather well optimized and although I state that it may be possible to boost the bubble on different frames for different directions, it isn't likely going to save much time (if it even exists, it hasn't been tested). Kid Niki (Ganso!! Yanchamaru) in 8:49, received 7.4 technical rating. I would give it a 9.5 because this game has been glitch hunted to death by me. I have not come up with any techniques or glitches. When inside a wall you can only zip upwards, not horizontally. The only known improvement is in the first level at a floating block that gives many coins (I believe that's the place I have in mind), because the jump onto that block can be improved by very few frames - otherwise I think of the movie to be perfected. Kirby's Dream Land in 8:44, received 8.7 technical rating. The TAS was well optimized on lag reduction. However, more lag reduction may be possible (it isn't very well understood why the same area sometimes lags more, and sometimes less). Subpixel optimization wasn't done at all. Dedede was rushed, so the strategy for him could be even faster. I think I would give this a 8.0 for those reasons. Volley Fire in 9:31, received 6.5 technical rating. I would give this a 9.0. Maybe even a 9.5, because I have optimized this very well and there are no particular known improvements other than "more optimizing". I don't think that many frames can be squeezed out of this though. Super Mario Land in 12:08, received 8.7 technical rating. It can be improved maybe by up to 1 second by using what I described in the forum topic for this game (delaying on the title screen etc.). Otherwise it is perfect. Someone said they have found a checkpoint bug which may improve this run greatly if it is true (which means I would give this a 5.0 rating). Unfortunately, the finder has yet done little to provide any information, raising the suspicion that his report was a fake (which wouldn't be unprecedented). Kirby & The Amazing Mirror, received 8.1 technical rating. I would give this a 9.5 because it was heavily optimized. No known improvements whatsoever (except for if we can manipulate CPU Kirbys to acquire Magic and use that to give us an ability mix whenever we want it - unfortunately, the RNG for that behavior hasn't been found yet. I have spent like 3 nights on researching it.) Tennis in 4:51, received 6.0 technical rating. I would give it a 9.0. There are no particular known improvements. Survival Kids in 27:58, received 7.4 technical rating. Maybe I would give it a 8.5. It was well optimized, and there are no particular known improvements aside from a console speedrunner that mentioned you don't have to listen to the radio (which would save about 1-3 seconds if it is true - I swear I tested for it and it didn't work though... Version difference hell?). Master Karateka in 2:39, received 5.3 technical rating. I would give it a 9.0, because it is well optimized with no particular known improvements. The final boss room is what keeps me from giving this 9.5, because the bird makes me walk (I can't run in presence of enemies), but it is probably unavoidable. --- Feel free to rate your own movies like I did. Or discuss movie ratings in general.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
I would like a revival of this. Maybe with shorter tasks, and maybe in a different format: How about giving people like 10 tasks and give them time until the end of year to complete them? Just throwing this out there. I'm thinking if I should host a competition like that.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
When drag dropping Super Mario Advance 4 - Super Mario Bros 3 (U) into Bizhawk 1.11.1 and dropping user movie #22117331642628592 from Kurabupengin's user movies, EmuHawk throws a fatal exception and needs to close. Also, that same message says in the end "depending on the error, it may or may succeed". This does not happen in 1.8.3.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
xNight wrote:
I don't know what i did but i was on 2-2 and i died in a area and it warped me to end of the level the the area where you go threw the door i don't know how it happened so I'm just gonna put it out there
source EDIT: It would appear that this game doesn't have an address for which checkpoint has been passed, at least I can't find it. So maybe it goes by something else (distance traveled through the level?) According to Frank Maggiore's RAM information, the checkpoints in each level are at "00, 28, 50, 78, A0, and C8". Although I don't know how to use these values yet, since I have not searched for a x-location address yet. FFE5 looks checkpoint related. (the screen I'm currently on. Imagine a level like a bunch of screens horizontally stacked together) It is more or less confirmed that the game goes by distance traveled rather than "you have passed this x location so now your checkpoint is ID xyz". If I go through a checkpoint and edit FFE5 lower again, I start at the checkpoint before that. So maybe the game was confused about FFE5 when the bug happened. I will try a few tests. FFE6 increments every few pixels. When it reaches 14(hex) it resets to zero and FFE5 increments by 1. The first checkpoint in 1-1 would thus be at a hex A0. ---- The only way I can see this bug work is if FFE5 can become 10 (hex) (the x position required in 2-2 to respawn near the end) in some way. Otherwise maybe it's some bug related to banks / cpu / stuff I don't understand well. Maybe FFE5 didn't get reset when 2-2 was started, as ais523 suggests. ---- When 2-2 starts for me, FFE5 and FFE6 read: 03 00 03 08 04 02 04 07 (this is the first frame you can pause on in the new level) 04 08 My question would be why it is 03 00 and then 04 02. Weird stuff gets written to those two addresses.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Thank you. Another question: How can I use free look in SMS without graphical problems?
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Just wanted to get a topic started on this. Maybe someone besides me will take on one of the games I made topics on here in the DOS section. ^^ I think this could make for a decent speedrun. In this click and point game, you play through a half dozen missions. There can be different scenarios depending on your actions (see this playthrough series that delves into that). I'm not sure if you have to restart a mission if you take a bad action. If not, it could make for a little fast paced run. E.g. instead of teleporting over to the hijacked ship and shooting people down, you can teleport a bomb over, get yelled at and just go back to your ship and take on the next mission. Since DOS TASing is still so user unfriendly, I don't know if I will be able to run this. (same with dozens other games btw, like Mystic Towers and Boston Bomb Club. At some point, someone has to guide me through JPC-rr again and really tell me from scratch what to do since I'm so lost. Or I wait until 2025 when another DOS emu comes.)
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Pinna park 1-7 TAS by 大福(zelpikukirby) Link to video I liked the pinna 6 void strat. Unfortunately the faster pinna 1 missiles weren't used (the thing I was talking about). I still think it should be looked into. You can save like 3 seconds with it. Btw, I'm probably not going to continue TASing the tasks, sorry.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
ThatEpicYoshi♣ on Youtube has commented on my glitch compilation video about a potential new glitch: as i got to the air port area i was messing around with mario and luigi's abilities (mainly fire-thunderhand) i was near the block, i actually hit the block a few inches away from the save album. then i ended up testing the elemental hand near the save album and suddleny it just started floating to the other block (that one was a wide M & L block.) then it stayed. i couldn't hit the save album because it was literally inside the wide one so i went to another area and back and it was in it's rightufl place. i'm not exactly sure what caused it. it might've been the wide block. idk, but this seems to be an un-discovered glitch. They said they used normal VBA 1.8.0-beta 3. Just posting this here for future reference.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Cartina wrote:
The Contra wrong warp needs to be documented and possibly used. I believe there is more to this. http://www.twitch.tv/dk28/c/5686445
Unlike the Rockman 2 thing, is this legit?
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Is this the character John de Lancie was known for prior to MLP? So the things Discord does is actually a play on this.
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Biospark, if you have better maps, I would like to see them. I know there is a hidden room before Level 4 boss, but it's useless. There is also an awesome clip I found which will save a few frames. http://i.imgur.com/FOINB7T.gif
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Some big guy on Youtube made a glitched run and got 80k views with it. (which is probably more than all other TASes and speedruns that ever used that glitch combined, which is funny)
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
! = good dump C = gameboy color, at least I think
Editor, Experienced Forum User, Published Author, Expert player (2316)
Joined: 5/15/2007
Posts: 3856
Location: Germany
Ball clip Link to video Would have been useful in Level 4 if the boss wasn't so super slow to kill without goku blasts. I don't know yet if ball item exists in Level 5 and if it can be done there. @BioSpark, those maps you were talking about would be useful now^^ Edit: I have found maps on vgmaps. Sorry, but I appreciate your offer!