Posts for Darkman425


1 2
9 10
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Hmm. I took a look at the RTA times for arcade Mortal Kombat II on SRC. Going by RTA timing rules of starting from selecting the character to when Shao Khan is defeated in round 2, this TAS has an RTA time of about 7:14. Despite using a lower difficulty compared to the RTA rules, this is slower than both arcade speed run categories for Mortal Kombat II regardless of the usage of the high punch trick. HPT RTA record at this time: 5:56.433 No HPT RTA record at this time: 5:59.633 According to the submission notes of the No HPT RTA record, they theorize that with perfect luck the time can reach the 5:4X mark. Those times beat your TAS by over a minute with more optimized strategies. I recommend taking a look at the RTA strategies and try another go at this game for a faster time that beats those times. It's worth taking your time to make sure you've pushed the limits of the game as much as you can.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Took a quick visual check. The green shell kicked by the kicker koopa just before the moon room pipe in Pekora Planet is definitely grabbable with a little bit of extra time taken. A look at the gap seems like it should be possible to get a shell jump across. With that situation specifically, it's a question of whether doing a traditional midair shell jump is faster/more feasible than having the shell ride the 1F0 blocks and bouncing off the shell at that point instead. Either way would cut out the wait on the goomba recovery to cross the gap.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Learning how to optimize sections is more of an acquired skill with different possible processes. There isn't any silver bullet method that works for everyone since different TASers have different methods on what to look for. That being said... Since there are readily available RTA individual record videos for the courses of the DS version of Sonic & SEGA All-Stars Racing, you should definitely be comparing the lap times between your work and the known records available. If the RTA time is faster your TAS time isn't close, then it's worth taking a closer look at those videos and see exactly what's going on there. It's also worth reposting the info I managed to find out about item RNG and how it can be manipulated, from my judging notes:
Darkman425, from the judging notes of the previous submission wrote:
One of the main things I had to check was if it was possible to RNG manipulate what the item capsules give. With a little bit of work, I figured out that this is determined by the in-game timer when the item capsule is touched. I managed to change what I got by switching between braking and accelerating on different frames so the item capsule is touched at a different time. With that knowledge in hand, it's possible to reasonably avoid getting hit and losing speed in the parts of the TAS where the player gets hit.
This does mean that there are possibilities where going slower can counter-intuitively save time on the total race. This does mean you might have to widen the scope to beating total race times rather than individual lap times. The core idea of what might work is the same since you luckily have readily available RTA times for comparison: focus on just one course at a time and keep refining it until it at least meets or beats known times. Of course this does usually mean spending a lot of your time to actually do all that research and production but you have to remember TASing is something that takes time and experience to become good at.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I had a small discussion with another judge before making the branch name change. The site has a tag for publications that corrupt memory values rather than having it as the name of the movie branch. As the game doesn't immediately jump to the ending sequence directly, it's not quite "game end glitch" so I went with "warp glitch". It can also be called "save glitch" as the process to jump to the end involves messing with event flag values and I'm still not super solid on what the branch name should be. I would like some more feedback on a proper branch name. I'll be asking some more staff on what works for the branch name, and it would definitely help if there's more feedback here as well.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I should post my WIP for those who are somehow still interested in making TASes for this game. User movie #638424460885615278 It's definitely not perfect considering that I've been manually manipulating RNG. There's even a couple of gaffes mentioned in the user file info that'd require an extensive amount of redoing RNG that I don't quite have the energy do redo. I'll probably step away from this for a bit but will probably come back to it. If nothing else this does give an idea of how an "any%"/"all gold medals" TAS on the US version could look like since the Pachisi glitch is abused to proceed to sequence breaking. There's even a bit of extra luck manipulation to get a poison knife from the Pachisi track, though it might not be necessary for an any% TAS. Anything that skips the pyramid would probably be attractive for the folks interested in TASing this version of the game as well.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I came across this thread on GameFAQs doing a bit more research since the Pachisi glitch grants access to the Open spell. https://gamefaqs.gamespot.com/boards/450388-dragon-warrior-iii/80368559 Turns out there are some doors that the spell can't open for some reason. More importantly, the New Town status can't advance without first obtaining the Final Key. While that does mean getting the dry vase is mandatory, the Pyramid and the magic key is completely optional. Alongside that info, the monsters in the Pyramid can be found elsewhere so even for all gold medals it can be skipped. This leads to a rough start of what can be done on the US version in terms of routing:
  1. After meeting the King, buy a warp wing and recruit needed party members.
  2. Head to Reeve to set a warp point.
  3. Get the Thief's key from Najimi Tower and warp back to Reeve.
  4. Get the magic bomb in Reeve and leave the island.
  5. Enter Romaly at night to grab the two tokens. Nighttime is required as otherwise a dog sits on top of a token.
  6. Head to the first Pachisi track north of Romaly. Grab the token on the 1st floor and the token on the board proper.
  7. Perform the Pachisi experience bug for each party member to reach end game levels.
  8. Head to Shampane Tower and beat Kandar 1. This is required to spawn Kandar 2.
  9. Warp back to Romaly and head towards Portoga. The gates can be passed with the Open spell. Talk to the king of Portoga to get the letter for Norud.
  10. Walk back to the Romaly Gate to reach Samano's Cave. Get the Mirror in the cave.
  11. Enter Samano Castle at night and defeat the Boss Troll, retrieving the Change Staff for story progression.
  12. Warp to Romaly and go through to Norud's Tunnel.
  13. Head to Baharata Cave and defeat Kandar 2. This makes the Pepper available in Baharata.
  14. Warp back to Portoga and trade in the Pepper for the ship.
  15. Take the ship to Edinbear. Use the Invisible spell and solve the puzzle for the Dry Vase.
  16. Head to the location where the Dry Vase is needed and get the Final Key,
This route skips over needing the Magic key in the Pyramid as a high level mage gives access to both the Open and Invisible spells. I haven't quite finalized where the rest of the route should go in either any% or all medals but taking a look at RTA routes can probably help with that.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I've been messing about with an all gold monster medals TAS for this game. Part of the routing involves figuring out how exactly the Pachisi bug works. I think I've got it understood well enough to get the bug to work consistently with some RNG manipulation. This thread on GameFAQs, specifically Paulygon's posts, was key to understanding how and why the glitch happens in the US version. https://gamefaqs.gamespot.com/boards/450388-dragon-warrior-iii/53836019 The mechanics guide on GameFAQs by Sk8erpunq was also key to understanding some of the inner mechanics. https://gamefaqs.gamespot.com/gbc/450388-dragon-warrior-iii/faqs/70252 Here's the rundown on how it works.
  1. The STR, AGI, VIT, INT, and LUCK stats are actually not integers. They have a shown value ranging from 0-255 and a partial value in memory. Each part takes up 1 byte in memory for a total of 2 bytes per stat.
  2. In the original JP release on the Pachisi board, max HP can be lowered by one of the events involving rolling dice to change stats. While max HP could be lowered, the current HP before the stat reduction still remains the same afterwards. This can result in having more current HP than max HP. For the US release, there was an added function to reduce current HP down to max HP after the stat reduction if the current HP would be above the max HP.
  3. However, this new bit of code has an oversight where the stat check function is done for every possible stat reduction in the dice stat change event. In these cases, the comparisons for most of the other results checks the partial values of stats in memory.
  4. For the case of the level up bug, the game checks if the partial STR stat against the partial AGI stat. If the partial STR stat is higher, then the whole STR value is copied into the high value for experience points.
This is controllable with RNG manipulation. RNG manipulation seems to work similarly to the Dragon Warrior Monsters game since they pretty much work on the same base engine. There are two important parts to this: the partial STR and part AGI values have to be set before attempting the Pachisi bug, and the shown STR value affects how much experience is gained by the Pachisi bug in this effect. Assuming a character has 65535 or less experience, this adds (65536 * shown STR value) to the character's experience. Levels ups are checked at the end of a victorious battle so the effect isn't immediate. Thus for a character to gain a lot of levels, the shown STR value must be a sufficiently high number to gain end/post-game levels. As a small aside, a sufficiently low enough shown STR value can lower a character's experience though this isn't exactly helpful. I have a basic RAM watch that I used for getting the conditions for the experience bug to work over here: User movie #638420079399910632
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Hmm, that seems like something that would slip my mind. A shell could be obtained with maybe a little bit of time wasted while waiting on the vine to grow. The shell might be better served taking out the first big mole to build P-speed faster and then taken to the upper level for clearing the other big mole out of the way, so that's a twofold potential place to save time.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Lots of new TASers caught my attention, so I'll split my nominations into two groups. Heavy hitters: JHolleworth Tuffcracker These folks came in with some really fantastic TASes out of the gate and I foresee that continuing from here. Specialized focus: DigitalDuck longbao Mikewillplays NemoRuby PearlASE SJ tom_mai78101 ZoroarkGX These folks brought in TASes that focus on a certain flavor to them. Whether it be game type, end goal, or system choice I can expect some good work from them. Big potential authors: Akiteru Captain_Forehead EnderOnryo McKniggediKnack Merl_ Here are some folks that to me seem like folks who could really bring some fantastic stuff in the future if they keep at it with TASing.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
[5309] PSX moon: Remix RPG adventure by poco_cpp in 2:04:15.18 Chances are that I'm the only one who's gonna nominate this one due to how few non-Japanese folks have heard of this game. It's a game close to my heart ever since I played the English Switch port, and I'm glad someone got around to making a TAS for the original release. It's also just a solid adventure game TAS with a lot of routing possibilities that had been researched.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Oh snap, someone found a death warp glitch on SRC: https://www.speedrun.com/master_of_darkness/forums/9uep1 Grabbing a health vial when at zero health via getting knocked back by an enemy attack results in recovering life, playing the death animation, and ending the current stage. This may or may not save time in all stages but there's definite potential in it.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
[5051] NES BlobQuest by Cephla, Samsara, ThunderAxe31 & Alyosha in 02:42.98 Watching new tricks getting found for this in real time was a pretty exciting experience. It's also a neat movie to watch as well. [5518] A2600 Xanthiom by nymx in 03:11.73 While nymx worked with me on the 100% the major zip glitches were found by him, and his research and knowledge in his solo effort of the any% TAS shine through in this one.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
[5059] GBA River City Ransom EX by tom_mai78101 in 08:48.03 The author worked on developing a pretty neat bot, the genetic algorithm bot, just to figure out how to get the best possible luck for a bunch of things in the TAS. Just just nuts to me. [5232] SNES Chrono Trigger by arandomgameTASer & keylie in 2:16:08.40 The careful amount of pre-planning and luck needed to improve a 9.5 year old movie still makes for a pretty tense watch for some of the major fights. [5399] GBC Pokémon: Crystal, Silver and Yellow Version "Coop Diploma" by gabraltar in 6:56:24.50 Having to handle 3 RPGs with and their RNG while also keeping it fast is incredibly impressive.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Dark_Noob His playaround TASes show off a lot of his creativity as well as presenting the games in a real neat manner. EZGames69 The run-n-gun TASes were pretty impressive, but also lmao you really did Big Kong. PearlASE Their background in arcade game research and knowledge definitely shines through in their TASes, especially when going for maximum score. Pepper-Color They made a good number of solid arcade TASes.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
[5216] N64 Blast Corps "all platinum medals" by WarHippy in 1:16:02.05 This game is notoriously difficult and it's satisfying seeing a TAS accomplish so many of these challenges with ease. [5395] Wii New Super Mario Bros. Wii "warps" by MGRoadkill, Squashh_1 & Qr1pt1k in 24:20.90 The whole power-up change up for the TAS route was something I wasn't expecting. It also led to a lot of crazy close calls which makes for an entertaining watch.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
[5605] NES Super Mario Bros. 3Mix "100%" by Mizumaririn in 2:07:17.64 I do like me 100% clears of hacks since they usually show off as much of what the hack is about as possible. A lot of creative levels done quickly, sometimes with unique tricks exclusive to the hack. [5619] NES Rockman 5: Air Sliding by CUI & NemoRuby in 25:32.96 I liked this one a lot when judging where I had to get some assistance for making a more watchable reduced flashing patch. A hack that gives plenty of powerful movement tools makes for a fantastic TAS watch for those who like Mega Man hacks.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
[5106] Arcade Street Fighter II: The World Warrior "playaround" by Dark Noob & SDR in 17:16.45 It's a miracle that original Street Fighter II could hold up to all this silliness. [5269] Arcade Raiden Fighters "maximum score" by PearlASE in 26:56.12 This is a nice and mesmerizing watch if you like some crazy movement in a shmup. [5715] Arcade Crime City by WarHippy, Lightmopp, g0goTBC, enderpal7, CasualPokePlayer, Invariel, Denial140, Truncated, RetroEdit, DrD2k9 & ThunderAxe31 in 06:55.67 I guess this might be kinda biased since I hosted the event that resulted in this? It's a nice and short platformer that's fun to watch and with some neat tech.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
The 2 Player 1 Controller mode can be considered its own unique challenge since it's basically doing a 2 player run but half the buttons aren't available. It would lead to its own optimization challenges with regards to movement options and enemy manipulations compared to being able to fully control both characters at once independently.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Hello, Blazingmo6. I am currently judging your submission but I have a lot of questions to ask for more information, especially since the submission notes are incredibly sparse. The races are all done on Easy difficulty. What does difficulty change, exactly? I do know that these games don't use rubber banding so I assume the racers would simply be more aggressive and more focused on their fastest route. My assumption is that due to the nature of how a lot of these races go, difficulty doesn't matter so much in races but leads to going past all the unlocks at the end of cups. Those unlock messages wouldn't count against you if you were aiming for in-game time rather than real time, of course. From my understanding, the warp lap skip glitch (or whatever it's supposed to be named) requires voiding out while in mid-air after crossing the starting line at race start? I noticed that you don't seem to get the speed booster item as often from item capsules. Is it possible to manipulate their rolls for more useful items without major slowdown? Speaking of the speed booster item, I also noticed the drift boosting speed trick going on. Would the speed booster item give better benefits, whether from making the drift boost faster if it's possible or just giving extra speed boosts when it's not possible to get a drift boost? I noticed that some turns weren't done as close to the turn as possible in a few races. Would it be possible that slowing down a little to get closer to the turn to save a little bit of time? There were a number of races with some parts that I found kind of questionable one way or another:
  • Whale Lagoon: I noticed that you drive in reverse rather than going forward the wrong way but turning enough once in a while to not void out. Is going in reverse faster overall in this case?
  • Shibuya Downtown: You get hit by an opponent right at the start of the second lap. That seems to be something that can be avoidable altogether without losing speed. I also happen to notice you're also next to road speed booster to your right when you get hit. If you had to get hit, wouldn't it make sense to take the hit and then be able to use that road speed booster to quickly get back to speed?
  • Treetops: You once again got hit midway through lap 1. Would a shield have worked to avoid that hit?
  • Jump Parade: You stopped near the start of the race before advancing. Was this to avoid the other racers or to setup the warp lap skip glitch? You also got hit right near the end of the first lap, which again I'm pretty sure is avoidable.
  • Pinball Highway: You bump into Sonic's car when turning around to go backwards, making you lose a little bit of speed. Is there a way to avoid this collision?
  • Sewer Scrapes: As with Whale Lagoon, you drive in reverse rather than face the wrong way for a bit but also you hit the turn a few times doing so.
  • Lost Palace: You bump into the tree before the second void out turning to the left. I also noticed the later void outs are done turning right where there's no tree in the way. Is it not possible to also do that for the second void out, avoiding the tree, or is that simply not possible to pull of the glitch as quickly?
  • Sun Fair: You bump into two racers when trying to go in reverse at the start of the race. Wouldn't it be possible to avoid them by slightly turning a bit during the moment you go forwards so the start of your reverse would be without interruptions and thus not lose speed?
  • Deadly Route: I noticed you slowing down right before getting an item capsule with a speed booster. This implies to me that there's at least one way to manipulate the item RNG but I don't know if it's by waiting frames, waiting seconds, or it's influenced by player inputs. At the same time I noticed you drove into another item capsule. Was there any possiblility that could have a speed booster as well without waiting?
  • Bingo Party: It seems like there was an extra void out performing the lap skip glitch. Was that one a failed attempt at getting the glitch to work that was left in?
My current feeling on this submission is that it could use a lot more time to refine and optimize this further.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I assume this could be a case where the first and second verification movies have different sync settings, such as different controller setups, and require being done in order to make a final SRAM for whatever needs it as the starting point.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I noticed the difficulty was set to the easiest difficulty. This isn't mentioned or explained in the submission notes. What does that change in terms of gameplay?
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Bay Route already has a 2 player publication but I can see someone wanting to see a solo run. [2332] Arcade Bay Route "2 players" by NitroGenesis in 10:31.72
Switch friend code: SW-2632-3851-3712
Post subject: DTC 11 Final Results
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Final Results First of all, it took a bit of learning AviSynth, VirtualDub, and trial-and-error, but I managed to make a basic encode showing all the teams' input files at once. Team 6 also provided an extended input file that was used for the encode. Link to video Reminder to self: upload contest entries to user files later, add neatly formatted spreadsheet comparison images. Here are the times, from first to last input needed to reach the ending. Participation Trophy 6th Place: Team 1 - 07:04.10 - User movie #638348116583870965 5th Place: Team 5 - 07:01.18 - User movie #638348122873732169 4th Place: Team 6 - 07:00.83 - User movie #638348124071385178 Bronze Trophy 3rd Place: Team 3 - 06:57.92 - User movie #638348119116981353 Silver Trophy 2rd Place: Team 2 - 06:56.90 - User movie #638348117734457371 Gold Trophy 1rd Place: Team 4 - 06:56.03 - User movie #638348121401609504 Congratulations to g0goTBC, Lightmopp, WarHippy, and enderpal7 for the winning entry! Also a special thanks to all the participants for joining in on this contest, and the previous hosts for advice. Comparison table Here's a spreadsheet of the cumulative time by segment. Here's a spreadsheet of the time to complete each section, with a sum of best to the side as well. Post-Script This was a fun and interesting learning experience for me. While I have been a GM for a mystery tournament before, actually being the showrunner was a bit more complex. There were a few unexpected bumps here and there but in the end it ended up being pretty smooth. Some games I considered but from the advice I received I had to cut:
  • Gunner's Heaven/Rapid Reload (PSX): It's a game that can be described as "We got Gunstar Heroes on PlayStation". Not used due to complexity from being that description and RTA runs being north of 30 minutes which would be a bit too long.
  • RoboCop (Arcade): This would've been my perfect pick if I hadn't found an offsite TAS out there. The most frustrating thing about that discovery was that the TAS I found was still slower than the SDA record.
  • Alien Syndrome (Arcade): Playing this game for speed would've been an interesting routing challenge, though I might not ask for doing a full 2 loop clear. This wasn't picked since there's already a published TAS of the SMS version that could've influenced some strategies despite being quite different.
Switch friend code: SW-2632-3851-3712
Post subject: Submissions are now CLOSED!
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
Submissions are now CLOSED! Thank you for all your submissions and your participation. Give me time to review and evaluate all the submissions.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1031)
Joined: 9/19/2021
Posts: 241
Location: Texas
I assume Peach is supposed to represent Mayuri if this is based on Steins;Gate. Reading the readme.txt file that came with the patch hints at the Konami code having other effects if used more than once. Do you know what that does, exactly?
Switch friend code: SW-2632-3851-3712
1 2
9 10