"Does not aim for fastest time, even as a secondary goal" pretty much instantly dooms this for the site.
This can be considered a form of low% speedrun; the first goal is low% above all else, but the second goal has to be speed. If it'd take, let's say, 12 hours of building up speed somewhere to save an A press, then go for it, but if you know you could save it in 6 hours instead, it has to be done.
For an encoding point of view, it's better to have the credits, and we can cut out the save your progress bit afterwards (like what is done for for Harry Potter PSX)
I rather have the encode be extended with a note that the last input has been changed to show the credits rather than extend the displayed time by ~10% of the run. I have poor experience with people who seem to only look at the displayed time on Youtube at a glance, and go "your run is slower than X" without actually knowing why it was longer. Sorry.
He's saying to make last input be what takes you to credits (which would make the displayed movie length shorter) and the encoder could do the save input (or you provide a secondary movie file that does the save input for encoders to use). I did the same with Grunty's Revenge. And I do agree it'd be nice to have the credits in.
I am completely unbiased in voting yes. It's definitely not because I've ran this game and my name is in the notes or anything!!
Here's a temporary encode I made to watch it myself, it should be done uploading in 2 hours.
Link to video
In my opinion, this is an underrated gem from Rare, and the best of their offerings on the GBA. It combines some great, tight 2D platforming with the ability to place animal buddies in the map with a variety of effects, which reminds me a bit of Lemmings.
The generous hitboxes make a lot of close calls look really impressive, I love the use of taking hits to save time, and the gameplay is at a nice pace to watch IMO. I suppose it might benefit from some glitches/tricks over the 50 minutes, but unfortunately it's a pretty solidly made game.
The difficulty curve is pretty nice, so the levels do get more treacherous, and the close calls more exciting, as the worlds go on, especially in Nightshade Mining Company I felt. And I loved the way that yellow ghost chased you in the final level, he was determined!
I love the (almost) complete lack of RNG in this game, because it means that everything done in this TAS is doable in real-time moreso than most TASes are. It makes it very relatable, at least to me...
Thanks for choosing this game to TAS, jlun, it looks like you did a fantastic job as always!
Ah, I forgot about this, I'll be sure to watch the submission and see if I notice anything that I feel could be faster (since improving levels is relatively pain-free). Right now I'll just look through your notes here...
A Blubba to get the Phoenix bag? That sounds fun and dangerous, good thinking to save menu and placement time.
I'm not sure I ever thought to use Phoenixes to prevent hurt animations, that sounds neat!
I could try to do the despawn glitch at more locations, but, I'd assume you already tried pretty thoroughly where it would be useful; the info I gave about the Stranglehold Swamp one is the most I know about it, seeming to require a pixel-perfect Sabreman location as well as letting the camera fully finish panning.
Yeah, using puddles as often as possible is good (it's especially neat how you can touch a mid-air one from below and get the speedup still, if I recall)
I don't have any memory of stores stocking different things randomly, but it's been a long time since I cared (esp. since I never did a non-any% speedrun). My only guess is, Blackwhyce is town 1, right? Was the Sucker shop from a state later on? Perhaps the stock changes when you get later in the game. I seem to recall shops sharing stock as well (Like, store 2 might be selling the same Serpent as store 1, but drop Blubba for something else). Just faint memories here though.
I know it's 1:30 faster by all measures that matter, but hey, that's why I said my motivation is silly! I wanna beat the time it's listed as, even though it's a completely meaningless, inaccurate thing! And TASing this is gonna be way more fun without grinding checking each frame...
From Blazephlozard's comments at the publication video:
Blazephlozard wrote:
It's all good, the horrid tutorial followed by menuing too fast for anyone to understand isn't cut out for moon I'm going to be doing another one actually, I'm writing a Lua script now that simulates RNG/shuffling to automatically check what frames might work much more thoroughly than I did manually, I'm hoping to go down to 4 deck edits
I might be able to help you with that. Back in the day, I had a C program (not a Lua script) that simulated the RNG. Unfortunately, I have formatted my hard drive several times since then, and I don't have it anymore, but it's really easy to find out what the game is doing. I see you have modified my Lua script, just use an emulator with decent debugging tools like BGB and set some breakpoints at the addresses it's reading, and you should be able to read the asm instructions and deduce everything. I might be able to redo this, it's not that hard, and the code not very long.
Anyway, I eventually gave up because I was busy with other stuff and I could not figure out how many calls to the RNG the AI was making in the battle, but if you just want to improve your search, doing that is not really necessary. Before I gave up, I could consistently find frames to do crazy stuff like finding 20 consecutive heads on coin flips.
Also, playing a bit of devil's advocate here, the movie was originally accepted before Vault, and was demoted to it just later, and it was not clear what criteria were applied to do that. I really appreciate the work you've put to optimize this game, I thought my submission would be rejected seven years ago, please continue working and don't worry about publication standards. There's a significant speedrunning community for this game, and if the movie is relevant to this community, it's all that matters :)
I've got the script working already, with help from Aroymart (RNG/shuffle code) and Mountebank (leader cheat criteria). It'll especially be a huge help for leaders, and makes it so I can very quickly check multiple different paths instead of going with the first one that seems good. These links should work (you have to run it when the decks are visually being shuffled; that's when the deck contents are loaded, but the order is unshuffled)
https://cdn.discordapp.com/attachments/93936126816956416/329221900612141058/TCG1_4_Prize_Duels.luahttps://cdn.discordapp.com/attachments/93936126816956416/329221906182045697/TCG1_6_Prize_Duels.lua
I don't care about vault when it's something like this, it's just not an enjoyable TAS to the common public no doubt, you have to know the game to follow what's going on at all, plus the damn tutorial! Same goes for speedruns of this game really, if you know the game, they're great to watch, otherwise, it's just a rapid menuing mess!
Right now I'm motivated by, as silly as it is, saving 10 seconds so that I beat your VBA real-time. If I have a better Water Club, better leaders, and save 1-2 deck edits, I think I can save that much.
Nice improving the former TAS (can't believe it only took you 13791 rerecords).
I did give this a thought that making a TAS of this game could be (mostly) automated in some way. However, no matter how easy MrWint makes it look, setting up automation for a game is not even remotely easy.
I mean, it's only 15 minutes, and probably a good 5 or more of that is just animations, booster pack opening, overworld movement... I had to completely resync a lot of the run each time I modified an old deck edit, but it's fairly trivial with TAStudio to visually see that the menu input in this game is frame-perfect. Insert a frame here, delete a frame there, get everything back on the first green frame.
Automation would be cool for that, but I dunno, I wouldn't trust it. I used a simple A+B alternator for mashing during the tutorial, that's the most I'd ever automate this kind of stuff.
What would be fantastic is some sort of script that can preview what hand a leader will cheat to, so I don't have to manually find it, as well as saying what the next coin flip will be. Those two things would speed up finding a good frame on leaders drastically. That might be fairly trivial, as long as the criteria for what shuffles they accept is set right. I might try and get something like that going if I ever redo this.
deuxhero wrote:
Any desire to do the sequel (which doesn't have a forced tutorial)
I have desire, and have thought plenty about it, but I'm extremely on the fence about it. One big con is that it'd be in Japanese, and being much lesser-known, it would have much less viewer interest. And being in Japanese wouldn't do me any favors either.
The most interesting thing is being able to get the Legendary Zapdos early from the Black Box; its Peal of Thunder Pokemon Power does 30 damage to a random Pokemon. This (and the new trainer Digger) are the only ways to do damage on your turn besides attacking, and would enable turn 1 kills in the many 2-pokemon duels.
But getting that specific Zapdos in your hand every duel you need it, plus enough Scoop Ups to re-use the power, plus manipulating the RNG of which Pokemon he hits... There's also all of the special rules, especially energy restrictions, which would make it very hard to minimize deck edits. Plus there's the decks you get for free which adds so many more options, as well as the access to the Deck Machine to save/load your decks which gives you even more options... All of this makes it way, way more complicated than this TAS, which only has one deck, and has a very simple strategy: Get a 20 damage attacker, exploit weaknesses, PlusPower as need be.
And without any benchmark to compare it to, I'm just not sure I'm up to doing it well enough to feel that it's optimized as much as it could be. I think I'd sooner redo this and hopefully beat VBA's frame count.
Funnily enough though, thanks to changes in when RNG advances in TCG2, you can actually do a real-time manip route. It relies on frame-perfectly selecting your file (which is a 2 or 3 frame window thanks to lag frames, so it's not too hard). You don't have the 60 options a second that TAS does, so there's much less 1 or 2 turn kills, but it's pretty nice: http://www.speedrun.com/run/oy27vnwz
Ahhh, you're who commented on my run <3 Great to have such an experienced TASer working on this!
I see you already noticed that jumping constantly is fast (I felt like that was the case but it's not very viable real-time except in swamps), and wanting to slide closer to the tent, and of course collecting the item as far left as possible (but that's not always for the best), that's the basics I had in mind for TASing.
The Bloater animal buddy seems like it might be kind of fast as well, it floats you upward and can temporarily give you movement speed much faster than running (followed by movement speed much slower). Not sure if there's a good one in the way to grab, though. That's the only animal buddy I remember thinking "this could be faster for some levels in a TAS"
The weird way you can despawn the fan enemies sometimes is definitely something to look more into of course!
One thing I remember that's worth knowing about and isn't really documented is how the Phoenix works; I think you still have normal invuln frames start whenever you touch an enemy with it. This is important for tornados; you can't walk through tornados with a Phoenix on unless you're currently "invulnerable". This can also matter for getting a bit more invulnerability after it ends.
I'd also look for more time-saving oddities in overworld movement, like walking against this slope here: https://youtu.be/OP1mKN2_cDI?t=2112
Did you try despawning the blower on Stranglehold Swamp (final level of world 3)? I tried and was able to despawn him, but not sure yet if it can be faster than using a boomer if done right since it seems to be a lot more picky than the one I do in the run.
The magic X position seems to be 834, then I have to wait until the camera fully pans over. The other 3 numbers in this are camera-related, the blower is despawned if I move right from here:
https://i.gyazo.com/0fe4470a88c071556241314ec7816125.png
It's odd, this doesn't KILL the blower, it only despawns it; no coin comes out, and it'll still be alive if you go left and come back. Unlike the one used in the real-time run, which full on kills them.
Aside from that nothing stands out as being improvable, except possibly for a few of the wolf chases where you slowed down, since there's many many ways to avoid the wulf (as qq mentioned).
Looks like a really good pick for a first TAS, and kudos for submitting it for April Fools. I hope you make another, of something a bit more exciting :P
This is good for Vault though!
Yeah, it's a meh from me, the bosses look very nice but the majority of the gameplay is just holding right and occasionally pressing B. total vault tier
It's awesome to see an open-ended game like this get its time cut down with a completely new route, and it looks like incredibly optimized gameplay all throughout too.
Hey Hyper, I hope you can see this message now that the chat has died down. I wanted to say congrats! I know you can do a little bitter but I always want to tip my hat off to you. Through thick and thin I will still be here to always support you. I know times are hard, but you always come back with a good attitude. I wish you the very best on your next TAS, and hope you get an even better score. I will continue to be here and find new strats as always. Great job Stiv. I am rooting for you always! You should be proud.
I thought of that at the time, and couldn't get it to work. And now I'm trying even harder now and still can't get it to work! The candle just doesn't travel fast enough to make it happen, at least for me... If you've got a video or input file for it, that'd be great.
I downloaded your movie and went to the part of the level to make a quick clip showcasing the trick. It turns out the jump was different than what I remembered. You have to jump across the pool, then turn on angle towards the ledge once halfway. I hope this works:
http://tasvideos.org/userfiles/info/27694957803882506
If that doesn't work, i'll have to make youtube vid to show it off :P
----------
Something else I discovered when playing through the game was that when in Crag Mill, you can skip dropping down through the fence below to reach the second level. There is a sac lying on the floor that is at a slightly higher elevation. You can jump off of it and make it to the 2nd level pretty easily.
Idk if it could be used to save time seeing how the 3 notes below would be skipped over. I'm assuming the note route would need to be changed for this level for it save any time at all, if it even could. But anyways, that's really all I can remember from my time with this game. I found a way to skip Talon Trot for low% but that's not too important. Well,I hope I've helped! :)
Ah, yup yup, that works! It looks like it saves 14 frames! And also would desync everything! Woo.
I didn't know that sack was that high, but even if I did, at the time, I definitely would have said, "nah, I need those notes". However, there actually are three notes that are probably faster. I timed the 3 outside Freezing Furnace lobby to be about a 150 frame detour on the way to Spiller's (and that timing is before I knew how to super jump), and I bet that's longer than it takes to go through that floor! Yay!
Wow, I'm glad to see somebody actually finished a TAS of this game! I started a TAS of this myself but I really only searched for glitches. I agree, this game is really underrated and one of the few last good Rareware games.
I probably should have posted this in the games thread, but I found a minor timesaver a while ago. When in the grabbing the orange Jinjo in the sewer as the candle, you can jump from the very corner of the edge and make it to the other side. It's really precise and could be done here to save a bit of time.
From what I remember, you need to jump at around that spot as the red line. Then, when more than halfway across the pool, turn up. You need to land inside the blue box. If you notice, the blue box is in the acid but there is actually invisible ground there mainly due to how the floor tiling works. Now, I'm not sure if this can be hex edited in or not (might mess with RNG), but it's not a big enough timesaver anyways, so no reason to lose sleep over it. :P Just thought I should post about it!
I thought of that at the time, and couldn't get it to work. And now I'm trying even harder now and still can't get it to work! The candle just doesn't travel fast enough to make it happen, at least for me... If you've got a video or input file for it, that'd be great.
OK, thanks, glad to bring attention to that bug~
It was a bit confusing since that's never how GBA TASes have started.
But it's better this way, since it's more accurately from power on of an actual console.
blazephlozard, It says it's forced to false for movie recording. Therefore, you don't have a choice. So what are you worried about? Who cares what the default is? It's been chosen for gamers, not tasers.
Well, but that's the thing. It's not forced to false. It says it will be, but it's not. You can record and playback a GBA movie in BizHawk, skipping the BIOS intro.
And every single GBA TAS I've ever seen doesn't start with the Game Boy intro, so, the precedent to skip the BIOS is there.
So I'm just asking, should a GBA TAS made using BizHawk have that option set to true or false?
Oh boy, I just started going through the intro of a GBA game earlier today! Guess I'll port that over, or just start over.
Should GBA TASes start from the Game Boy start-up screen now then? Based on what the Skip BIOS setting says, they should. Yet, Skip BIOS defaults to true...
There's a weird little thing right now too, where if you have Skip BIOS set to true, and start to record a movie, the recording starts from the game start-up, not BIOS start-up. (The Skip BIOS setting says it'll be forced to false for movie recording, but I guess it isn't.) And if you play back that movie, and do a Play from Beginning, it'll do the BIOS start-up first, which obviously desyncs everything.
So, the Fez PC port has came out.
http://www.youtube.com/watch?feature=player_embedded&v=Houl_emUdrkhttp://www.gog.com/gamecard/fez
As soon as it was announced, I wanted a 32-cube run of it to be my first TAS. Not high enough quality to submit, but more just a fun thing to attempt. Fez on XBLA is my personal game of the year 2012, so it'd be really awesome to do.
Problem is, I'm guessing it'll be nigh-impossible to get it to work.
The game is written in MonoGame, originally in XNA, and as far as I know none of the games I've seen working in Hourglass use either of those. When trying to run the game, no window ever appears, and FEZ.exe just hangs in the task manager at a 4.3K~ memory usage doing nothing. There doesn't appear to be any glaring errors in the log. The DLLs all get loaded twice and then the second copy is unloaded, might that be a problem? I don't know, I'm totally lost on how to jury-rig a game to work in Hourglass. Nothing I tweak makes any change.
At this point, people are having quite a few issues getting it to run in normal Windows, so, it'd be a miracle to get anything in Hourglass working! If anyone could figure out how to get it to work, I'd love you forever!!
Why? There's no restrictions for any other supported system. People TAS various versions of NES/SNES games all the time. Hell, the current Zelda 64: OOT TASes (both!) use the japanese 1.0 release, which is the first of eight versions (3 JP, 2 EU, 3 US).
Having eight versions (quite a few of them identical gameplay-wise) is nothing compared to the typical PC game. And, like I said, the huge issue I see is using one specific release, besides the first release, because it has the most bugs. They use 1.0 in OOT because it has the most glitches, as expected. When it's a glitch not in the previous version, and not in the next version, it's a lot different, kind of like you're playing a broken version of the game, a mistake. It should be in a category of its own. TAS is about breaking a game, not playing an already broken game, yeah? And this is an issue that's really only come up in this era of rapid patching. None of it really applies to this run, but, I think these questions are going to have to be answered for the future of Windows TAS, and I don't think the answer should be allowing any version at all.
Rydian wrote:
Last I checked you can't TAS the 360. :P TASes are created with emulators and game images, where it's no problem to feed the emulator a specific image. Even when game images can't be played on an emulator, they're still dumped almost as soon as they come out or update (see the 360 scene releases), and even when there's no way to play them people are dumping them (see the 3DS releases).
Obviously you can't, but I'd hope in 10 years you can! TAS can't be about N64 and below forever! If every version/image of 360 games are being dumped somewhere, that's great for the future. I was worried they weren't, and there'd be no way to get them anymore.
The whole issue of 2.0 vs. 2.1 here is very important to the future of PC (and newer consoles) TAS I think, as brought up before. Especially this quote:
boct1584 wrote:
I think the issue is that this bug was not in previous versions, and is not in future versions.
The only two releases of a game that are obviously acceptable are the first version, and the current version.
Imagine v1.0.6 of a game introduces some crazy collision bug that lets you walk through many walls and skip large parts of many levels. This wasn't in v1.0.5, and it was fixed in v1.0.7. Does it really make sense for 1.0.6 to be the accepted version for TASing, just because it's the one that can be run the fastest? Even though it was an official patch release, it wasn't really a "stable" release. But then how would you define "stable release"?
In the case of VVVVVV, and this run, it's absolutely not an issue, for a few reasons:
1. VVVVVV 2.0 is the first version of the C++ port, so in many ways it's like a 1.0 release.
2. Being "2.0", obviously it is a stable release, not an intermediate patch
3. The version of the game you download from Steam is still 2.0, so the 2.1 linked on the distractionware forums is not even the main commercial version.
But, I think it's important to think about the rules on what version numbers of a frequently-patched game are allowed. And in the case of consoles, could you even get the in-between versions? I know on 360 you can play the version on the disc, or patch it to the current version. It's not like every version of a patched game is stored somewhere. It's different than, "This cartridge I'm holding is 1.0. This other cartridge is 1.1." Patches overwrite the old game.
I don't know. It's something to think about, because I think PC TAS has a bright future. Hopefully this isn't gibberish.
Idk if it could be used to save time seeing how the 3 notes below would be skipped over. I'm assuming the note route would need to be changed for this level for it save any time at all, if it even could. But anyways, that's really all I can remember from my time with this game. I found a way to skip Talon Trot for low% but that's not too important. Well,I hope I've helped! :)
From what I remember, you need to jump at around that spot as the red line. Then, when more than halfway across the pool, turn up. You need to land inside the blue box. If you notice, the blue box is in the acid but there is actually invisible ground there mainly due to how the floor tiling works. Now, I'm not sure if this can be hex edited in or not (might mess with RNG), but it's not a big enough timesaver anyways, so no reason to lose sleep over it. :P Just thought I should post about it!
