Posts for MESHUGGAH

MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
This TAS's reckless driving part can be improved by around 22 frames (frame 5293) edit: I have lost my WIP which gets to this position but I believe the notes are enough to replicate the improvement. The improvement can be done by a better breaking to our only driving location. This changes: - the length of the cutscene (and it's relevant lag frames before and after IIRC) of going inside the building - the whole ghost RNG but these can be adjusted with door opens and closing, wasting only 2-3 frames per doors. Remaining notes I have with more details: * 3 frame saved by leaving the delaying of "A" on the 24 floor building above the HQ. Seems to be there is a frame rule regarding the transition between the "walking to the house" cutscene and loading the house. * 1 frame saved by fixing a mistake at the 18F (removed an extra A button) * 4 frames saved at the final battle with Gozer by manipulating ;Input processing The first input on a non-lag frame won't process it's defined action as long as you don't release it, with driving sequences and Gozer being an exception. So we always start from the 2nd frame most of the time. ;Lag patterns There's a few patterns that can be used to ensure the input being optimized: 11111111 - No lag 10010101 - Map screen 10101010 - 30fps (for driving and Gozer) The final battle lags very much while being 30fps. However, the time differences between shooting and registering the hit can be used instead for checking optimality. ;Road sequences The distance you need to go through the driving sequences increases with the steps done on the world map. Since we are already use the shortest path, only fine tuning is possible by pressing different (mostly directional) buttons when selecting a level. ;Ghost manipulation Can be manipulated by position and by opening doors. Not all doors can be opened for only a few frame, but the 5F -> 12 -> 18F is the fewest (3 x 2 frame) sacrifice can be made to go through all the 24 floors without dying to the ghosts in the process. ;Gozer manipulation ** delay minion's shots ** avoiding Gozer's shot two times ** minimizing and delaying the number of ghosts Moar details - Gozer has 16 hit points - We have - The shortest time to shoot Gozer is 22 frames - This movie is 98 frames behind the hypothetically shortest fight
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Woohoo! Will definitely start testing this and compare my notes with the quantium website notes. edit: I've skimmed through the 1st pages of posts and started updating my old notes. What is the preferred method of showing my test results? Asking this because I don't see the problems mentioned in this thread being reflected in the github issues. Current state of my notes (not updated fully, yellow means not checked for new site) to demonstrate what I would show instead of creating N number of new issues: edit: I have removed the sheet as I have found at least 2 issues that could be abused. Please PM me if you want to check out my sheet! Also, the search... Holy moly it is lightning fast!!! edit2: (Note, I'm just an old tester of this site) no design document? Now I know why there are many inconsistencies abroad pages. While the structure of tasvideos content probably won't change only the number of data to be parsed, there should be a standard. Otherwise who and how does he/she agreed upon different visual changes? Replying to the most frequent problems: - yes, there should be a scroll to top of the page (not represented in github issues) - colorblindness and dark mode: use DarkReader as mentioned by someone already in this thread or other CSS changing extensions. I prefer to use DarkReader which automatically changes the colors and you can manually configure it with sliders and options and save it to use your own style whenever you visit that domain. - "width of content area" yes, I also see this as a huge problem: the overall content of the pages got scattered and way more distances than previously. It became the opposite of being compact.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
I've had and lost my WIP which improved the previously published TAS by a huge margin up to the 3rd floor (1st floor we arrive after jumping in the 1st window from top). So, here's my observations: - Suicides seems to be optimized (fastest already) - Jumbo's 1st level probably optimized - Rooftop rappeling probably optimized Now for the hard part... 1. The terrorists have a "moving" priority with first moving between rooms and secondly moving between floors. There is no straightforward strategy apart from a manipulation viewpoint. The movements place is unknown (AFAIK they only move on non lag frames) 2. The game has a whole array of those terrorists, so the counts are actually legit. Each terrorist is a real entity and can be parsed in lua to get an interactive map. If the game would be easy, strategy would be something like this TAS: eliminate current floor, descend to next floor. The game is not easy and you can get scenarios like the terrorist moving up the floor while you already descended. You also need to time entering doors and rooms at the right time when the terrorist already there to avoid waiting time to kill them. This means you have the following edit: factors to build up your strategies: - Try killing everyone on top floor, while trying to manipulate terrorists to go bottom - There are 16 room per floors, terrorists can be anywhere even on stairs - Terrorists usually don't traverse stairs up, but there is a very little chance of.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
According to many year old notes of mine, this TAS can be improved at the very beginning of the 1st level by 1 frame. I only have "optimized movement" as details, but I remember the powerup should be within this time save, so the improvement is located between jumping on the first pipe and picking up the powerup.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
This TAS loses 1 frame at the very beginning because Start and Select spread over two different frames instead of using it in a single frame. This leads to a huge desync because the game is a real mess. But hey, the potential is there.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Interesting case. So the TAS syncs on console using FCEUX RAM start pattern but doesn't with the originally used BizHawk RAM start pattern? This means only the output/result of the TAS is verified, while the starting/initial setup used not. I'm not sure if "partial" console verified TASes should also get the console verified tag. Since the TAS itself doesn't syncs only after changing the required RAM pattern.
Alyosha wrote:
My NES is not the most reliable, so this process can take a couple tries.
So after using the clear cart, power off, edit: monopoly cart, power on, the TAS doesn't syncs each time? Are you sure it's about your NES and not an unknown variable/factor somewhere else? Full list of potential factors - your NES as you mentioned, - your RAM clearing/changing cart, - your Monopoly cart, - the time spent between ejecting your ram clear cart and powering on monopoly cart (you should aim for sub 10 seconds. Greatly depends on platforms, obviously the faster you are the better your chances will be. In case you are desperate, you could try https://en.wikipedia.org/wiki/Cold_boot_attack ) (no more edits)
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
It's not a problem if we don't work on this that actively, I'm still trying to figure out how does the spawn system works. - It looks like they are based on cam x poses and they can be avoided by waiting ~18 frames. - Enemy direction before spawning depends on cam x (Kenshiro+camera) pos - Since only 2 enemy slot exists, this leads to N spawn positions / levels where most of them are blocked/avoided by running constantly forward while the 2 slots are being used - 2nd enemies don't always spawn, just because it's slot is free. Currently my WIP6 level 99 timer for first level is 1387. Ideally we would just spawn 2 enemies right before entering a boss fight, so we wouldn't need to alter any other enemies. However they spawn and might be way too much behind us to let an enemy spawn in front of us, or even two.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
After a lot of thinking and internal (me with myself) debates, I'm going to modify my original project goals and it's contents in the following ways. The main change comes from the following: - Avoid any possibility of potential cause of TASVideos.org site negative future (legal law cases leading to closing or financial punishment) - Avoid any possibility of potential cause of a member of TASVideos.org site negative future (anything and any kind of personal details with no exception, anything legal, anything illegal) This leads to the following changes: 1. Avoid direct linking to ROMs and copyrighted files - Since the point of my project is to ensure the common knowledge between TASers are on the same level, there will be either links or guide on possible sources of obtaining them - These kind of links should only exists outside of TASVideos site - These links will include already known links that TASVideos site uses (for the record I'm not going to disclose which is this site despite it's name and their project goal is well known) - There might be files that only I possess, have no idea how to include them without interfering with TASVideos' no linking to copyright stuffs rule. The possibility of this and the number of files are really really low, so this is just going to be revised after a complete version of the project materializes. 2. Avoid enclosing any kind of TAS related personal details - This means I will drop researches that tries to assume or find out what a TASer meant for example in forum threads. An extensive research could lead to TASers deleting posts to avoid representing themselves in historical events, which is definitely not a goal for anyone including myself. - This doesn't affects researches like developers own comments and facts about creation of video games and their bugs, but I'm going to omit / redact personal life details for example personal cases, anything that doesn't directly affected the potential causes in the viewpoint of developing and it's life path. For example Cyberpunk 2077, it should include every detail except the name of legitimate sources. It should include everything from death threats to lies and corruption. 3. Incorporate my researches to TASVideos Wiki - As suggested by feos, I believe this should be the right way. We have an incredible but very hard to navigate Wiki full of worthy information. Duplicating these would defeat the purpose of a central knowledge base. - I need to come up and check if it's possible to make a Wiki page with the ability to order tables or use google drive as a host. The main question is the balance between easy to read, easy to add or modify information, easy to check the authors. 4. Reducing my main project into an introductionary booklet - As of the above point, I want to make sure that the original contents of TASVideos pages remain intact while demonstrating different viewpoints on introducing different kind of users to the world of TASing - The introductionary means a gist of all stuffs related to TASing and TASVideos without elaborating them to atomic level. The point is gathering my viewpoints and applying the same guidelines for every topic that comes up in the community. - Depending on what comes out of this, they will be allowed to be copied or use words to be pasted to TASVideos pages, which I'm going to avoid. This is simply because I don't want to "spread" my own words and it should be done by others especially in the staff to ensure the historical texts of TASVideos while enriching them with my own findings or my conclusion of researches. Any kind of feedback is welcome as always!
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Mission accomplished!
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
I've made my researches and the USA should be used because of those long cutscenes in the Japanese version. I will finish polishing my Lua script tomorrow (Saturday) and try to get an optimal as possible version of Stage 1 with the least amount of position/time sacrifices to get double explosion glitch for both miniboss and main boss. And also will write down Alyosha and Kyman's TAS to compare us against it with correct, precise comparison points.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
According to the extensive researches done by Decentralized Radical Autonomous Search Team Investigating COVID-19, the actions made purposely or accidentally while experimenting with a destructive stuff is definitely was and will always be a concern affecting everyone on the world. It doesn't matter who does this. What matters is the knowledge of anyone doing it to be able to react in time in case a disaster something like coronavirus happens. Or Chernobyl. Or nukes. Or anything that's possible to experiment with without others discovering it. While I used to joke about should we just hang out in hazmat suits, I hope that we don't progress into wearing an all hole covering diaper suits.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Wow! That was way fast to create a 3:17:50.07 [709348 frames] TAS. Obligatory good job! According to the submission text it's 9 edit: and 6 seconds BIOS time) seconds slower than previous TAS because of better emulation. So the math is 59.727500569606 rounded * (9 + 6) = (9+6) *60 = 540 + 360 = 900 frames. The differences between versions seems to indicate that your jpn TAS is faster than tas_suki's: 737 frame difference, 163 frames faster. edit: Sorry for not understanding your sentence originally. At this point, it would be better to know the exact duration difference than using an assumed time difference estimation. Frame counts: 1070926 - currently published TAS 1071823 - this submission 709348 - alternate version for jpn by (Jigwally &) CasualPokePlayer 708611 - nicovideo jpn tas by tas_suki
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Thank you for the elaborated answers taters! Yes, this was the video I was missing to understand this TASGiving as someone who heard it for the first time. Well, depending on the video game and it's TAS popularity, difficulty and it's technical knowledge, questions might arise in form of twitch chat users' message or as a generic interview between the staff and the author or commentator who submitted: - viewers might have questions to the author (how did you come up with this idea, what motivated you, why are you so good, is this RTA viable) - viewers might require more time or knowledge to understand a specific run (requesting elaboration or analogy, drawings or showing webpages with information, speedrun being way too fast to be commentated) - viewers might have information (while this is usually only true for streams with relatively low count of participants to be able to share a possible improvement or fixing a mistake while the commentator speaks) As a viewer, I always like to know as much as possible about the runners and their mindset, their stories and history of finding out the solution to problems, the crazy goals and attempts done in order to achieve 1st place. These are usually conducted with an interviewer who has knowledge of the games and able to ask questions the commentator is able to answer. edit: these are just ideas based on what I've seen on GDQs, but I've also loved the ability for the chat to participate in manipulating the events of playing a game. There's a lot @ Wiki: TASBot (not every tasbot appearance are documented on the page if I'm correct)
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
I was following the majority of GDQs which is hosted on twitch chat, were (usually) physically based public event with microphone announcer, all kind of guests, scripts for an actual production value entertainment, interviews and other charity and donating related stuffs. I wasn't following the speedrun scene between (edited) 2019-09 ~ 2021~07 apart from sometime glancing at the reddit speedrun. I've read the site news text, this thread OP post and your post. I'm still unsure if I understand what is this event or which video from dwangoAC should I watch to understand what is happening (as a production, not the games themselves). edit: twitch chat interactivity was part of the TAS block at some GDQs
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Haven't updated the code since my last post in this thread. The original (and still planned) idea was: - automating downloading and formatting every entry from speedrun.com and putting it to a new page. - redone SDATASes the same way apart from 1 obstacle The reasons for this are the following: - SDA runs were mostly submitted under strict rules, so all (or at least most) speedruns should be a good idea for TASing - SDA is... somewhat... dead if I'm correct. This is good for me, because I could (and plan to) somehow gather around all SDA runs and check all TASes related. I see only 1 obstacle like some runs listing every platform while only 1 platform has that run. - speedrun.com runs are mostly submitted under community or single person made up rules, so the majority of speedruns doesn't have enough strict rules or have way unique rules that might be not a good idea for TASing. - for the same reason, there might be corrupt/cheated/fake speedruns as well - speedrun.com runs data is not uniform. There are runs where only IL records are present, runs with multiple branches, runs with different timings... My future plan would be consists something the following: - SDATASes recheck first, because it's a good resource for probably good run ideas. I've used c# last time but I think I will hop on the python train, gather all runs, find a way to identify original platform the run uses, output similar to what we already have on SDATASes page - edit: for speedruns.com, reconsider what type of runs and what games should I analyze. I guess it's not that bad idea to have all speedrun.com runs with no TAS to be assembled on one page, but it will definitely have a note about the possibility of the runs having corruption.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Thanks for the answer! No ACE/memory corruption for a non any% branch: Ahh okay, this makes sense! So I believe using ABSTUDLR to end the level is OK. Regardless of what is actually happening in the background, being a "warp to next level" or simply "end level" as this is reached by doing the jetski glitch and manipulating memory values and other stuffs directly and indirectly, skipping only 1 section of 1 level among many levels). I guess I will look into that somewhere in the future, not because of the bounty that I'm not even interested in but I think I could come up with something. Oh there were some nice discoveries around that linked post, will read them sometime :)
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Oh my god, yes vote!
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Since I guess this TAS will end up in vault, I think you should use the Japanese version. 4:58:59.01 versus 3:17:44.065 is definitely a huge difference.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
I'm out of the loop. Can I request for some more information? Since the 12 page long form can't be skipped until required fields filled in. I guess it's possible to submit commentary speedruns online (live stream) and or offline (recorded)... ? I opened the following links but couldn't figure it out. https://tasgiving.org/ --> no video https://docs.google.com/forms/d/e/1FAIpQLSc3ed4NX0Dk4_vAFI4WOl2q4eC1TlatWOYbGDQKBPRYdpOTwA/viewform?usp=send_form --> volunteer form with required fields https://docs.google.com/forms/d/e/1FAIpQLSdLLWBVmLpBLT1fxHljfqGSnl7QnIGYZae07YDfzW3RWAFJFw/viewform?fbzx=-6871505502756041612 --> submit form with required fields edit: okay I see on dwangoAC's youtube channel the previous TASGiving's submissions where some people with webcamera or not commentate on their TAS, there doesn't seems to be any interactivity between the viewers or any host... right? edit2: err so these are real time speedrunning stuff with live comments? judging by this video: https://www.youtube.com/watch?v=H0Lft38iK6Q
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
While I did not watch this TAS, I've already watched the beginning of the first TAS incarnation of this game. 5 hours of sokoban or even 20 minutes of sokoban is not entertaining for me. Duration difference usually doesn't change my perception of entertainment of a TAS, especially if that so slight. edit: so what's up with the improvement called as "questionable"? According to the observation from the submission notes, using the probably next version of BizHawk that emulates better than the current one is okay, as long as that version going to be released where the TAS will sync.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Okay, just wanted to let you know. Just realized I should have asked this through PM to respect your privacy, I'm glad my mistake was "okayish" now.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
5 hours of sokoban. Voted no.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
According to my notes (my WIP has been lost through mistakenly formatting my hard drive), the first level of the previous TAS can be improved by 17 frames with optimal speed controlling over obstacles 3 times. This TAS doesn't improves it that much and this mistake (which I actually already written on the youtube encode 5 years earlier) edit: can be seen when jumping over tree obstacles. Approximate times of 1st lag frame at the end of stage 1 Rocky Forest Randil 3900 This TAS 3889 My lost WIP 3883 edit2: a question for the author: are you aware that your youtube channel name doesn't reflecst your username change here at tasvideos? edit3: and the submission file as well
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Good job! Yes we need to change the comparison points to be more precise. Level timer 99 My latest hud script here: User movie #73353016102155940 (will upload and add to this post soon) uses gui.pixelText(120,8,string.format("%d%d", memory.readbyte(0x65), memory.readbyte(0x64))) 0x65 the first digit of timer for example 9, 0x64 the second digit of timer for example 9, 8, 7, 6... you know. Powerup address being the enemy ID Okay, this makes sense, will do some tests today. Yes, main bosses seems to be always F and have different gameplay than normal minibosses which seems to be normal enemies. The rest of the values (past F) seems to be the result of how the game works. Imagine something like the ROM having the following regions [enemy data] [level data] [sprites] where using more than F probably reads from another region. edit: For example the enemy data might only contains 15 entries, so reading from enemy data 16 would read from the offset (position) over the end of it's region, reading from another region. Times Updated the timesheet a lil bit with your post' numbers. WIP5 which uses a different kind of position sacrifices to get a little better 2 enemy position before 1st miniboss for double exploding boss glitch User movie #73353048780342013 edit 1 for this post: I will continue researching on Tuesday or Wednesday. Note that while it's good to have a movie with so many (frames) improvements than this submission, the movie will probably desync, so we will use -- your new WIP -- movie as comparison points and when we are about to continue from 1st miniboss end. (As soon as we can come up with a way to verify the potentially best movement/time sacrifices to manipulate the position of 2 enemies before starting a boss fight), we can add entertainment relevant stuffs as well as trying to improve each part of your newer WIP. final... edit: Our new strategy so far (Applies to both Hokuto no Ken 2 and Fist of the North Star, but we are probably going to use the USA version. Will start with checking up the version differences and edit the gameresource page) Stage 1 (level 0) - Manipulate two enemies as close as possible to the first mini boss cam x position 1792 - First mini boss double explode - Manipulate two enemies as close as possible to the final boss cam x position 3840 - Final boss double explode Stage 2 (level 1) - This level has 3 stairs, which increases our cam x position greatly: (cam x position in previous room, cam x position in next room) from 1070 to 1560 from 1794 to 2304 from 2200 to 2816 - This level might be improvable with using damage boosting if possible. This depends on how far we need to scroll the camera and how many time sacrifice required for projectiles and/or enemies boosting us as fast as possible. - Manipulate two enemies as close as possible to the first mini boss cam x position 3328 - First mini boss double explode, pick up glitched next level power ups for Stage 4 (level 3) Stage 4 (level 3) - Going back to the left to start boss fight and double ko The remaining stages seems to be all rely on the previously mentioned strategies. Boss fights where enemies don't spawn will use double ko, while others will do double exploding glitch.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
MESHUGGAH
Other
Experienced Forum User, Published Author, Skilled player (1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Currently I believe the powerups also functions as other visual stuffs, considering that the 0x70 is the "open a door" animation used in the later stages as screen transitioning. More addresses with explanations: (later they will be added to the gameresource page) Main boss hp 0x6a --> seems to always load 56 for every main boss Mini boss state 0x6b --> Animation state (Door) 0x72 Animation state (Main boss, enemy slot 0) 0x78 EA, EB: garbled enemy sprite, no effect EC: first enemy (from stage 1, level 0) sprite ED, EE: No sprite, but doesn't gets removed on its own EF: garbled enemy sprite, no effect F0: Kenshiro lying (dead), no effect F1: garbled enemy sprite, no effect F2: Kenshiro standing, no effect F3: Main boss #1 F4: garbled enemy sprite, no effect edit: F5: Kenshiro ducking punching, no effect F6: nothing, gets removed fast F7: garbled enemy sprite, no effect F8: nothing, gets removed fast F9: No sprite, but doesn't gets removed on its own FA: Kenshiro standing kicking, unpickable power up FB: Garbled Kenshiro, no effect FC, FD, FF: No sprite, but doesn't gets removed on its own FE: Garbled boss sprite, no effect Regarding the snippet of code: Line 1: X = 0x8A's value == 5 Line 2: X = X - 1 == 4 Line 3: 0xEE = 0xEA + X == 0 I'm not sure if I'm right about the last line, will revisit a bit later
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...