1 2
11 12 13
Samsara
She/They
Senior Judge, Site Admin, Expert player (2241)
Joined: 11/13/2006
Posts: 2823
Location: Northern California
TheWinslinator wrote:
If you don't have bizhawk set up, this makes me wonder why the transitions between levels in your tas are much shorter than mine. Which emulator and what settings did you use for it, if you remember?
Looks like Gens was used for the published runs. BizHawk's much more accurate with loading and lag emulation, that's why the transitions seem slower. We don't count this as losing time for the purposes of judgement (as it's unavoidable when getting more accurate emulation), so there's no need to worry about it. Now that I think about it, you might actually want to hold off until the next version of BizHawk is released. If I remember correctly, there are a lot of updates and changes coming to the Genesis core (something huge is happening with it at least), so it's highly unlikely that any work you do now will continue to sync on the next version. Of course, you could just finish it all on 1.11.6 and be completely fine and not have to worry, but just in case you feel like the run's going to take a while...
TASvideos Admin and acting Senior Judge 💙 Currently unable to dedicate a lot of time to the site, taking care of family. Now infrequently posting on Bluesky
warmCabin wrote:
You shouldn't need a degree in computer science to get into this hobby.
Active player (487)
Joined: 2/19/2007
Posts: 114
Yes, Gens was the emulator used for the runs I made, but that was because it was the only option available when I made those runs. I would recommend sticking with Bizhawk as it is a more accurate emulator and looks like it has more features as well. Many games on many systems have longer transitions when we move to using newer more accurate emulators from older ones, so don't worry about it. I've recorded a short section of BLW1 on BizHawk, so you can see more easily if you are really loosing time against me or not, though I didn't thoroughly compare it to my old movie and I am a bit rusty, so it's possible I dropped a frame somewhere. A couple of places are a bit ugly too, as I didn't take time to smooth out the movement, but that shouldn't effect anything from a comparison point of view. I've uploaded it along with a very simple ram watch file that you can load in BizHawks ram watch tool, though you may have already been using that. http://tasvideos.org/userfiles/user/TheRandomPie_IV The ram watch contains only the horizontal position and horizontal speed, split into pixel and sub-pixel components, as I couldn't find my old file from when I made my runs, but those are the main ones you need to be able to properly see what is going on and properly compare two attempts at a section. If you've already improved your first attempt a bit this might not be that useful, but if there's something else you want to know fire away.
Site Admin, Skilled player (1255)
Joined: 4/17/2010
Posts: 11495
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
Samsara wrote:
Now that I think about it, you might actually want to hold off until the next version of BizHawk is released. If I remember correctly, there are a lot of updates and changes coming to the Genesis core (something huge is happening with it at least), so it's highly unlikely that any work you do now will continue to sync on the next version. Of course, you could just finish it all on 1.11.6 and be completely fine and not have to worry, but just in case you feel like the run's going to take a while...
The core itself is not getting improved, it just moves to x64 and starts using a different approach in savestates (while core app state gets saved now instead of what its author decided to put to savestates). This improves stability and determinism, so accuracy-wise it's the same thing. And there's no need to wait, because no one knows then the release is out.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Alright I'll just work at my own pace with this then, should be done with this probably in less than a month. RandomPie, I actually haven't been using RAM watch, as I'm still learning the various features of the program, and had no idea I could access such useful information just by putting in their hex addresses. I've already shaved off a bunch of pixels just using these variables. This should be all I need for an Any% run, but if I find I need another variable, such as Y-position/speed, how easy is it to find such an address? Do you use the program XVI32 to do that? I've also been toying with the idea of doing a Maximum Score TAS for this game, as there are things like max coins for Super Mario Bros. I've been able to get 1,817,001 in 2 hours RTA, and suspect a TAS could probably get 1,817,500 or so in 20 minutes. Also, we could finally show off helmet reset glitch in a TAS! Thoughts?
Patashu
He/Him
Joined: 10/2/2005
Posts: 4045
I approve of max score runs in general.
My Chiptune music, made in Famitracker: http://soundcloud.com/patashu My twitch. I stream mostly shmups & rhythm games http://twitch.tv/patashu My youtube, again shmups and rhythm games and misc stuff: http://youtube.com/user/patashu
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Alright, since I haven't found any other useful timesaves in a while, I'll go ahead and post two files here. The first file is the one I did pretty much all by myself. The second file contains some sections which have inputs spliced from RandomPie's 12:47 run. https://drive.google.com/file/d/0B5wBv6IivLIOTVhCUXdPeHVQRnM/view?usp=sharing https://drive.google.com/file/d/0B5wBv6IivLIOdzA3T0FEZVphMnc/view?usp=sharing In that file, inputs were attempted to be spliced starting at the first incline of BLW1 all the way to the Red Stealth helmet in BLW2. Due to some pixel misalignments, however, I was only able to do it from after the rubber block pit to the Red Stealth helmet. The spliced file saves 10 frames over mine, and probably another 5 if the inputs between the first incline and the rubber block pit are optimized. So RandomPie, I'd be willing to give you credit as a co-author if you're ok with me using the spliced file, and maybe you could clean the TAS up in a couple areas, as you've got way more knowledge than me on this game. Thanks to your RAM watch file, I was able to save a good number of frames on collecting the diamond at the beginning of BLW1. I also improved the Plethora ending a bit by taking a different route to the flag. Questions: Is there any reason why you hold "X" in the entirety of your TAS? Could my method of hitting the shooter blocks in BLW2 potentially be faster than the method you used? As of yet, mine is barely slower.
Active player (487)
Joined: 2/19/2007
Posts: 114
Nice, I'll try to check those out this weekend. I'm fine with co-authoring and taking inputs from my old runs, and I'll try to help out and go over things as much as I can. There's no reason for the X holding, I think I have X bound to the same key as one I'm using for something else and never bothered to change it because it has no effect.
Active player (487)
Joined: 2/19/2007
Posts: 114
Ok, so I got around to properly testing your method of handling the blocks at the start of BLW2 and it turns out that it does seem to be slightly faster (just under two full frames) than the method I used. Here's my best attempt so far. http://tasvideos.org/userfiles/info/33172480726180494 Nice find!
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Ah, that's great to hear! I could easily see one of the improvements you made was activating the two blocks then moving a slight bit to the left, so as to come back right with a higher speed than I would have had hugging the blocks until they disappeared. One thing that gets me about your TASing style is the seemingly random places that you do 1-frame left inputs. I've observed that it is faster to do that in the places it is done, but it is beyond me why it's faster. Also, how are you able to determine the precise time in the jump to hold left? It looks like arbitrary places to me, but I would guess lots of trial an error to obtain the optimal result.
Editor, Reviewer, Experienced player (980)
Joined: 4/17/2004
Posts: 3109
Location: Sweden
TheWinslinator wrote:
One thing that gets me about your TASing style is the seemingly random places that you do 1-frame left inputs. I've observed that it is faster to do that in the places it is done, but it is beyond me why it's faster. Also, how are you able to determine the precise time in the jump to hold left? It looks like arbitrary places to me, but I would guess lots of trial an error to obtain the optimal result.
It's not in random places, as I'm sure you have figured out. It has to do with the odd way that air drag is calculated. The Game Resources page for Kid Chameleon actually has an explanation of it.
In the air, all characters except SkyCutter decelerate 0:1 per frame. This is called air drag throughout this article. Air drag is present regardless of if you walk off an edge, jump, or bounce off enemies. If your speed goes below 2:95, it will instantly be set to 2:160.
http://tasvideos.org/GameResources/Genesis/KidChameleon.html At certain points it is faster to turn around, and decelerate on purpose to go below 2:95. This gets you back to max speed (2:160) faster than if you had let the speed tick down 1 subpixel for several frames. The average speed is higher even though you brake for a short while. If you play the movie frame by frame and check the speed (both whole pixel and subpixel) in the RAM watch RandomPie gave you, you can see it happening and probably figure it out.
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Oh yeah, I remember reading that information in the guide, actually, but never put two and two together. Seriously that is a genius way to gain time in a way few would have expected. I mean, I've been wondering that since I first watched the TAS back in 2011. Well, with that in mind, I may take a look over what I have so far once again and use this nice little trick in a couple places. Thanks for the explanation, Truncated!
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Alright so here's what our progress looks like so far. This file is 15 frames faster than the last one I linked! http://tasvideos.org/userfiles/info/33414609093679181 BLW1 - Incorporating the air drag reduction strats Truncated told me about, I was able to improve the level (not including the first section where the diamond is collected) to exactly the number of frames that the current TAS takes, plus two improvements: 1/2 frame saved due to a subpixel manipulation at frame 588, which avoids Kid having to walk up the next incline for one frame. 1 frame saved due to conserving bounce momentum a bit longer. You can see Kid lands on the platform at frame 1137 slightly farther to the right than he did previously. BLW2 - Spliced RandomPie's shooter block segment to the beginning for 1-2 frames saved. The rest of the frames are saved after collecting the Red Stealth helmet. At around frame 2450, we no longer land on the block the helmet came from, and we can move downwards much faster for a big timesave. A couple additional frames are saved thereafter through more efficient jumping. Also got a bit more creative with swordplay :P Plethora - All I found here was getting to the flag 4 frames faster using RTA timing, but really doesn't matter at all since the TAS time was unaffected.
Active player (487)
Joined: 2/19/2007
Posts: 114
Nice! I was a bit busy recently so I didn't see your question before, but it looks like Truncated had it covered. I made some attempts at BLW1 myself and ended up with this: http://tasvideos.org/userfiles/info/33462102174540930 It completes BLW1 in the same number of frames as your movie, but the improvements came at different places. I gained a little bit at the first dragon, but then fall behind at the first slopes, and then gain ground again entering the rubber block pit, but didn't improve the end section like you did. It looks like it should be possible to be 2/3 frames faster in BLW1 if we combine the best sections of both attempts.
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Ok here's the new bk2: http://tasvideos.org/userfiles/info/33481442694174900 This was very interesting. So, I spliced in your beginning dragon method, where you gained a lead by quite a few subpixels. Then, I put in my inclines segment, which put me 1+ pixels ahead of you. However, your rubber block pit is just way better than mine, due to the fact that your precise pixel position allows you to be able to tap down one less time than me to avoid "feeling" the rubber blocks. In order to implement this strat with my 1+pixel lead, I unfortunately had to reduce my lead to only ~50 subpixels; only then would it work. Then, implementing the longer jump at the end, the TAS was actually improved by a whole frame! Oddly enough, we're two frames ahead of the old video when the flag is captured, but the load time before BLW2 was one frame longer for some reason, hence only the single frame saved. That being said, what do you know about lag in this game? Not only between level transitions, but also for example between frames 855 and 856, Kid actually moves negative 191 subpixels. It loses over a frame, and I'm wondering if it's lag-caused. I tried everything to try and fix it, but it seems to be related to ducking itself in that spot.
Active player (487)
Joined: 2/19/2007
Posts: 114
The lag can be a bit weird unfortunately and I never really got to the bottom of it. I've tried a few things but I couldn't do anything about it either. I can't think of anything else that could improve BLW1 at the moment, time to focus more on the other levels I think.
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Sorry I couldn't reply earlier but I got loaded with work these past few days. Just wanted to let you know that I think I've found everything I can find, so I'd be ready to publish the TAS whenever you are. But if you're confident there's more left to find, by all means, keep looking for improvements and take your time. Ya can't rush art!
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
I decided to take a look at this again after a bit of creative inspiration this past week, and I was able to shave off 294 frames from the last movie! http://tasvideos.org/userfiles/info/36009974368461154 I cut 9 frames from BLW2 due to a better helmet collection, and the majority from just barely being able to pull off the bounce from the second to last tank to the top of the incline (made possible with better air drag reduction). More importantly, this saves an extra HP for Plethora. The rest came from a massive overhaul on the Plethora fight. Due to the extra HP, getting Micromax is no longer necessary, saving 112 frames from transformation time alone. I then experimented on several different methods with Red Stealth, the best outcome actually being far away from the flag, due to the arrangement of obstacles between that area and the flag. In the previous movie, I had to jump off a platform to the flag 75 frames after Plethora was defeated, but now the last input is made 70 frames BEFORE Plethora is defeated. Additionally, I found a funny side-effect of running off moving platforms while ducking every other frame, which teleports me instantly above Plethora, saving some frames (if you don't know why it happens or why it's useful, I could explain it if you'd like). I'm aware that the title of the bk2 says the time is 1:21.95, but when using the actual console framerate, it's actually 1:22.07. If I can get the latter time below 1:22 (-5 frames), I'd be satisfied with the outcome. But right now I'm at the point where new frames to save are really hard to come by, and I was wondering if you can offer any insight as to where further saves are possible - if you have the time of course. My .tasproj file still has branches of other Plethora methods I did that didn't end up being faster. If you want to experiment with those as well, I can link that file here too.
mklip2001
He/Him
Editor
Joined: 6/23/2009
Posts: 2227
Location: Georgia, USA
Hey, I just saw the new Kid Chameleon run put up on SDA, and it encouraged me to look for this forum. Are there any plans to submit this new Any% route? I don't know if there was firm consensus on it being called an in-game cheat... I think that run would be fine as a separate category on this site.
Used to be a frequent submissions commenter. My new computer has had some issues running emulators, so I've been here more sporadically. Still haven't gotten around to actually TASing yet... I was going to improve Kid Dracula for GB. It seems I was beaten to it, though, with a recent awesome run by Hetfield90 and StarvinStruthers. (http://tasvideos.org/2928M.html.) Thanks to goofydylan8 for running Gargoyle's Quest 2 because I mentioned the game! (http://tasvideos.org/2001M.html) Thanks to feos and MESHUGGAH for taking up runs of Duck Tales 2 because of my old signature! Thanks also to Samsara for finishing a Treasure Master run. From the submission comments:
Shoutouts and thanks to mklip2001 for arguably being the nicest and most supportive person on the forums.
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Yes it still is! See, Kid Chameleon will be having its 25th anniversary on May 28th, and I've been saving submitting this until then, just so I could guarantee that I have some content to put out on that day :P There have been four runs of this category submitted in the past, and it doesn't look like any have been rejected for the reason you mentioned (though I fear it will be a hot discussion point). Plus, this run has a significant advantage over the others, being over 13 seconds faster than the quickest of the four. Also as a side note, if you check out speedrun.com, I've already beaten that run by 2.3 seconds lol: http://www.speedrun.com/kidchameleon#Any
Joined: 2/19/2010
Posts: 248
Just wanna say I'm super excited about this news! Looking forward to the run :)
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
I'd just like to let everyone following this thread know that the TAS has been submitted! Go ahead and check it out: http://tasvideos.org/5524S.html
Post subject: Kid Chameleon (Genesis) Secret Warp Zone
XTREMAL93
He/Him
Active player (446)
Joined: 1/6/2012
Posts: 587
Location: Azerbaijan, Baku
At the end of Blue lake woods 2 there is a series of blocks just above the flag. Jump to the last one and then press Down, Right, Jump and Special together. You will be warped to Plethora, the last level. Link to video
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
We’re one step ahead of ya on that :P http://tasvideos.org/5524S.html
XTREMAL93
He/Him
Active player (446)
Joined: 1/6/2012
Posts: 587
Location: Azerbaijan, Baku
Winslinator wrote:
We’re one step ahead of ya on that :P http://tasvideos.org/5524S.html
yea i know) just want to document this video tutorials here
Experienced player (940)
Joined: 7/18/2016
Posts: 107
Location: United States
Get ready guys, some improvements to the currently published TASes are definitely on the horizon. A YouTube user by the name Favini01 has discovered and TASed some pretty impressive stuff recently, the latest being that airwalking into a boss head as Berzerker while in the charge state absolutely decimates the thing. Check out their quick work on Shishkaboss below, which I would estimate saves about 15 seconds over the current maximum levels TAS. I suggest you subscribe to this guy as I have confirmed they will soon be TASing similar for Bagel Brothers. I'm anxious to see if the time save there is just as impressive! Link to video They also found that Skycutter can in fact gain speed from bouncing off enemies if you face backwards when the bounce happens. Even though the video below is from a ROM hack, I've tested it and the speed gain does in fact happen in the base game as well... although I don't think the base game has any situations that could yield a speed gain this... extreme xD Link to video
1 2
11 12 13