Posts for Darkman425

1 2 3 4 5
9 10
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
For time comparisons, I assume loading times are why this initially seems slower than the Switch and Steam rereleases. Is there anything outside of that where those versions would be faster than the PlayStation version that I should also know about?
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
The limited user role as it is seems to be something that sticks with someone for months and years. I think the idea around the limited user role was to be more temporary and have said TASers try to take their time refining and understanding how to optimize. In practice the way the limited user role seems to drive folks away. I don't think the limited user role isn't necessarily a bad thing to have but its use as it is didn't seem to have an intended purpose. I remember an old conversation about how some users sometimes spam a whole bunch of submissions in a short time span. I feel like in those cases giving the limited user role temporarily to someone who does that would make sense, even if it's for a week or something. It'd at least discourage holding onto a number of complete projects and sending them in all at once. Even then those situations are incredibly rare and can also just be handled by regular moderation methods. For now I'd say the limited user role would best be suited for extreme situations. The problem would be defining what those situations would be as the only one I can think of at the moment is the hoarding problem as I mentioned above.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Typically Game & Watch games have a value that keeps track of game difficulty that goes up the longer the game goes. Does Parachute X Octopus really not have a value for that?
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Alright. I'll wait until proper submission text is up to make sure I understand what's going on to the fullest.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
I've watched the TAS in emulator and also looked up some of the speed techniques used in this. This seems like it covers what I think is most of the techniques that should be known to understand what's going on, including why carrier view isn't used: https://sites.google.com/site/blastcorpswiki/techniques Should I hold off on judgment until actual submission notes are added, or is it fine for me to go on ahead once I make time?
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
So about that. I had problems getting it to run from the merged set before I figured out that MAME checks for names rather than hashes. Wherever I found my merged version had the right file hashes but didn't have the names MAME was looking for to run bkonggx so I had to deal with renaming some files.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Alright, it's not April 1st anymore and I got time. This one was a bit tricky to get working due to some file shenanigans. If whatever version of bkonggx (the version used for this TAS) found doesn't seem to run in MAME, there are files in ckongg (unencrypted Galaxian hardware version of Crazy Kong) that share the same hashes of the missing files but not the file names MAME's looking for to run bigkonggx. These were the files that had to be taken and renamed from ckongg to bkonggx:
ckonggbkonggx
ck_cp.binmmi6331.6l
ckvid10.bingc.h1
ckvid7.bingc.k1
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Now that's what I call audio clipping!
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
An attack for every worm maggot noodle of bucatinni on the floor.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
When playing back the input file on BizHawk 2.8, I got a desync at around frame 56077 at the fight in the Gamer's turf, where the teleport goes to the wrong spot. The attached image is what happens immediately after the teleport. If it does sync with a released version of BizHawk 2.9 or you have the time to fix the desync, I can set this submission to Delayed. What I did watch that did sync was enjoyable to me, but I would like some more feedback on this run.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
I think the best way to check is to see the boss flags that are checked. This submission should have all the relevant bits to check: #8040: hellagels & klmz's GBA Castlevania: Aria of Sorrow "Julius mode, all bosses" in 06:13.04
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
I found a 89 frame improvement by simply skipping past the Sega screen at the start, and all of your inputs sync perfectly after that first adjustment. Are you fine with me replacing your input file with the faster start? I would not add myself as co-author for this minor adjustment.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
The MSXHawk core only supports MSX1 games as it is right now. Metal Gear and Metal Gear 2 are both MSX2 games and thus are unsupported in BizHawk.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
From watching it, I assume the "wrong warps" in question are the vine skip in 4-2 and using the pipe before the lava in the cheep cheep room of 8-4.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Your video link doesn't work since it's not from a video streaming site like YouTube. Please change the link if it's possible for you.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Hello, and welcome to TASVideos! There's a few emulator pages for running MAME games for the purposes for TASing: There's a lot of resources and support out there for working with LibTAS if you're willing to take the time for setup. Those can also apply to MAME TASing and lets you immediately work with whatever versions of MAME you need in case of regressions and such. For something in the future that's really promising, there's a MAMEHawk core being worked on for TASing arcade games in the multi-system emulator BizHawk. It's still experimental right now but will be in BizHawk 2.9 when that releases to try it out. Since it would be in BizHawk that means there are a number of tutorials for using BizHawk that can help. Once you got an emulator set up, the easiest way to get a handle on TASing anything is simply to jump in. Always be willing to ask questions on the forums and the TASVideos Discord if you have questions.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
A bit late to respond to this. I like the frenetic movement being done and I want to see what a completed run looks like. Nice work!
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Hello, Eduardo_EVF, and welcome to TASVideos! I'm here to conduct a review of this submission so here goes. Review: I got the movie to sync, but it should be noted that there's a lot of blank input after the last input needed for the ending sequence. Removing the blank input means the actual length of the input is around 11:43.38. I watched the current RTA world record to compare against your TAS to check some things: https://www.speedrun.com/dt2gb/run/zn8w359z Something I noticed throughout was the RTA record was faster than your TAS in small places throughout the game. When comparing level by level, there's small optimizations seen in the RTA run that slowly gets ahead of the TAS, whether it be doing lag reduction or better slight movements to reduce waiting. The language differences only apply between stages as the number of textboxes in stages remain the same, so version differences aren't a big factor. Something I also noticed was that on the Mu boss the RTA record seemed to get more favorable boss actions resulting in a faster kill. Luck manipulating the Mu boss to use the rock dropping attack would definitely make that fight faster. If you could get that luck manipulation to work there, it's possible that the other fights could also be sped up with better luck. It's a perfectly reasonable first submission at this game but it definitely needs a more fine tuned run. Take your time to first research strats if there are any available such as existing RTA runs out there. Try applying those strats in places your run seems slower in as well as try possibly faster strats if you catch something others didn't see. You have as much time as possible to make a TAS that's very refined and hard to beat. tl;dr A number of small places where time could be saved compared to RTA runs add up across the run, making this submission suboptimal.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
An alternate input file that has sound would be nice for alternate encodes but not necessary. There have been publications that forego sound and/or music to save time, most notably the Mortal Kombat 3 speed TAS: [4733] SNES Mortal Kombat 3 by KusogeMan in 04:51.27
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
So I was reminded of this game due to Discord. I just remembered that I saw a weird early exit glitch in an old race video from 2018. It could skip ahead a few levels. This plays the Super 2 Mario Bros. version of this bootleg, though they're the same game pretty much aside from some text and the main character sprites. Pay attention to the top left screen for the glitch in question: https://youtu.be/YG2ajOVoyeU?t=1525 Not sure if it would help a warps run but it's potentially worth looking into for anyone somehow interested in this.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
For Game Boy games running on Game Boy Color or Game Boy Advance mode some games can have their initial RAM state modified in a meaningful way by switching the initial color pallet in the BIOS splash screen and has been proven to be console accurate. Poking FRAMES seems to be similar to that with RAM being modified before starting the game. I can see that as simply setting something to manipulate RNG as long as it's also correct with real hardware. From my understanding, something like poking the FRAMES variable falls under that category.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
Zippy the Porcupine is a homebrew Atari 2600 game by Chris Spry and Atari Age that adapts Sonic the Hedgehog to the system. This game started as a physical release but has since been made freely downloadable by the developers here: https://forums.atariage.com/topic/269247-zippy-the-porcupine-binary-released/ I did a quick and not super optimized test TAS that I uploaded here: User movie #638114940436284980 I also made a preliminary RAM watch file uploaded here: User movie #638114942205017629 Some notes I made when making the test TAS:
  • Jump height is influenced by Zippy's speed when the button is pressed. This means for certain jumps Zippy needs to be at the right speed to get successful jumps.
  • The best way to maintain speed is to be running, so getting out of rolling on the ground is usually the best course of action outside of when it's needed to get through passages or destroying enemies that are unavoidable.
  • Spin dashing with down and the button creates a frame of lag. Since speed can be built up at the start of a stage and during vertical area transitions, spin dashing at the start is about a frame slower. However, spin dashing is faster by about a frame if there are places where Zippy needs to stop and start again, such as falling straight down to a lower level and rebuilding speed.
  • Vertical screen transitions have a load times and from my observations should only be done if there's no other way around.
  • The input polling is normally runs at 30 FPS but there are a lot of effects that can slow this down, such as landing, destroying enemies, ground breaking, etc. This is the worst with the Aqua Zone levels since that naturally has a lot of lag since those levels do input polling at 20 FPS when in the lower level for water physics simulation.
  • Boss fights are influenced by player inputs. This is highly important for the final boss since the times between hits can be varied by about 100-200 frames. All bosses take 6 hits and the fights end upon touching the emerald on the right side of the screen when it drops to the ground.
  • Boss locations, or at least the first one, is determined by the frame which the button is pressed on the title screen.
  • At the minimum, I got a level order that only required 10 out of 16 levels to be played.
I might come back to this but I know there are Sonic loremasters that could improve this a bit. The main thing I want to do is to maybe get a boss order that has as few Aqua Zone levels played as possible but I'm not entirely sure if all boss locations are determined on frame start or when each boss fight ends.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
When making any TAS it's always worth taking a fair amount of time and research to see if someone already has already got information and tech for your game. Sonic related content is generally popular enough that more likely than not you can find someone who's posted either RTA or TAS runs of a given Sonic hack. Finding those means you now have benchmarks that you can meet or beat, and even ask for help if you can't figure out how to match their times.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
A quick review of this one: I got sync to work for this homebrew game. The creator of the game has it available for download here: https://github.com/xram64/falling-nes I found one other possible way to manipulate RNG: the frame start is pressed at the title screen determines what the initial RNG seed for the run is. Seeing as this submission already is 35 minutes, any testing for a better initial seed should probably be handled by a bot since doing otherwise would be unreasonable. Otherwise, the TAS itself is fine as it achieves reaching the maximum score of 999 and ends input early without overflowing the score.
Switch friend code: SW-2632-3851-3712
Darkman425
He/They
Editor, Experienced Forum User, Judge, Published Author, Skilled player (1112)
Joined: 9/19/2021
Posts: 247
Location: Texas
A quick review: I got this to sync on my end. The final input needed to get the ending started is at 11:20.05, which makes this over 40 seconds faster than the publication done on the first loop through the use of the counterattack strategies on all the bosses. The final input in the submission is done after the ending to display the third loop's final message at the end.
Switch friend code: SW-2632-3851-3712
1 2 3 4 5
9 10