Survival Kids is a game about a boy or girl who gets trapped on an island and tries to survive and escape. There are multiple endings in the game.
This TAS abuses the text glitch (aka. hut glitch) to reach Ending 3 and complete the game almost three minutes faster than klmz' published "Ending 1" run.

Objectives

  • Uses a game restart sequence
  • Abuses glitches in the game
  • Genre: Adventure

About the run

This game first caught my attention when I watched klmz' Ending 1 run a long time ago. I liked the gameplay and the fact different endings could be reached so I took a closer look at it. Eventually I began work on what would later be the Ending 8 TAS. While working on that run, I discovered a text glitch that happens if you bypass the trigger in front of the hut and then enter the hut.
Basicly, the trigger makes the text pointer (and ROM bank) a certain value, which makes the character say certain text and go enter the hut. Once inside the hut, more text is displayed. The game doesn't adjust the text pointer when entering the hut, but instead it assumes the text pointer wouldn't have changed between the two cutscenes, in front of the hut and inside the hut. Since you can bypass the trigger by pausing and restarting the game, you can walk around freely, allowing you to change the text pointer by using items or checking stuff on the overworld before entering the hut. After doing so, a wrong (and possibly glitched) text will be shown based on the erroneous text pointer and ROM bank. This glitch was long known about and tested thoroughly. Spikeman - who has been a big help in understanding Super Mario Land 2 Pause glitch - even jumped aboard the glitch investigation but we didn't find any uses in any of the three versions (US, EU, JP).
About two weeks ago, Melonslice kindly pointed out the glitch can be abused in the Japanese version, where the game's day counter is set to a seemingly random value without the game crashing. This only works if you are playing as the girl, for some reason. This way, having the day counter at 99 or higher and sleeping will trigger ending 3 or 7 (depending on whether the glitch makes it so the "other kid" has been found). It is also possible to complete ending 1 in some cases, since the game is made to believe the helicopter is supposed to arrive. I began investigation very quickly, and put my results in a google docs sheet.
Unfortunately, the Japanese version doesn't let you skip the intro cinematics. More differences are documented at TCRF.
The fastest way to beat the game seemed to be to reach ending 3 by eating a grass that triggers the "ugh, bitter" text, setting the text pointer to 16550 and ROM bank to 5. This resulted in a 20 second long glitch text (unlike most other results, where the glitch text would last for 1~2 minutes or longer). The name was set to のとの for no particular reason, because a short time needed to be spent waiting for the correct food scenario (so the "ugh, bitter" grass appears in the right spot). I chose this name because the kanji looked funny to me.
I would have loved for the glitch to allow me to complete the same ending as klmz, ED 1, reasonably fast. But the glitch didn't cooperate. It seems the fastest way to go for ending 1 using the glitch would be to use the Match item before entering the hut, setting the text pointer to 23019 and ROM bank to 3. This can cause the day counter to be set to 2, so sleeping once more will cause the helicopter to arrive magically. But the glitch text in this case is running slowly and it takes about 2 minutes until you regain control. Thus, ending 1 can be completed in about 4 minutes.
I've been asking myself, should each ending have its own run?
I don't think so. Not every ending has enough merit. You could argue that the glitch run should be its own category, but otherwise I would say there should be "fastest ending", "best ending" and "all endings". That's how I see it, but I will let the judges decide.

Possible improvements?

Someone on the Youtube comments mentioned they had figured out the possibility to take over the Program Counter (PC) and maybe achieve ACE, but they forgot to post about and document their findings. Perhaps there will be the possibility of a faster or better glitched run (such as "ending 8" or "all endings" in 3 minutes) at some point.

Thanks

  • Melonslice, for pointing out the possibility of a glitched speedrun.

ThunderAxe31: File replaced with a movie resynced on BizHawk 2.2.2, and judging.
ThunderAxe31: The text in the ending cutscene got broken, but the ending still counts as cleared since the game sets the relative completion flag after movie end. Accepting as improvement over the currently published fastest completion run.
fsvgm777: Processing.


TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15581
Location: 127.0.0.1
This topic is for the purpose of discussing #5924: MUGG's GBC Survival Kids "game end glitch" in 02:45.70
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
It was short and sweet, and it didn't leave any feel of disappointment to me, so I guess I'm giving a yes vote. By the way, why didn't you include the BIOS screen? It's officially supported in the latest BizHawk release. It would probably be possible to easily resync it on it.
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Active player (261)
Joined: 12/13/2016
Posts: 352
Like Clock Tower, this should obsolete "any%" for fastest ending.
Post subject: Here's an avatar I don't get to use very often.
Skilled player (1416)
Joined: 10/27/2004
Posts: 1978
Location: Making an escape
I'm... sorry, but the girl reminds of of Ariel from Peter Gray's Draw Manga.
A hundred years from now, they will gaze upon my work and marvel at my skills but never know my name. And that will be good enough for me.
Editor, Expert player (2329)
Joined: 5/15/2007
Posts: 3933
Location: Germany
By the way, why didn't you include the BIOS screen? It's officially supported in the latest BizHawk release. It would probably be possible to easily resync it on it.
I wasn't aware there was an update to Gambatte. Would you like to look into resyncing it?
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
MUGG wrote:
By the way, why didn't you include the BIOS screen? It's officially supported in the latest BizHawk release. It would probably be possible to easily resync it on it.
I wasn't aware there was an update to Gambatte. Would you like to look into resyncing it?
Ok, I'll see what I can do. Edit: Done! It was trivial, all I had to do was adding 186 frames at beginning and change Equal Lenght Frames to False, otherwise the game skips a bunch of input polls. Please always use this setting for Game Boy TASing.
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Editor, Expert player (2329)
Joined: 5/15/2007
Posts: 3933
Location: Germany
Active player (261)
Joined: 12/13/2016
Posts: 352
Why not just make it the default setting?
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
MUGG wrote:
Thanks ThunderAxe!
No problem!
ruadath wrote:
Why not just make it the default setting?
That's a great idea, since we can't expect everyone to know about this detail before starting a new project with GBC BIOS. I've posted an issue on github: https://github.com/TASVideos/BizHawk/issues/1176
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Editor, Emulator Coder, Expert player (2156)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
ThunderAxe31 wrote:
By the way, why didn't you include the BIOS screen? It's officially supported in the latest BizHawk release. It would probably be possible to easily resync it on it.
It might be a bit off topic, but why should the BIOS screen be included? The BIOS screen is not part of the game software.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Skilled player (1741)
Joined: 9/17/2009
Posts: 4981
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
klmz wrote:
ThunderAxe31 wrote:
By the way, why didn't you include the BIOS screen? It's officially supported in the latest BizHawk release. It would probably be possible to easily resync it on it.
It might be a bit off topic, but why should the BIOS screen be included? The BIOS screen is not part of the game software.
Any runs that use a hard reset sequence would also be "faster" than a real time run, due to not having to play the BIOS. Of course, this run doesn't use it, but I would count personally for consistency to what happens on actual console.
Editor, Emulator Coder, Expert player (2156)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
jlun2 wrote:
klmz wrote:
ThunderAxe31 wrote:
By the way, why didn't you include the BIOS screen? It's officially supported in the latest BizHawk release. It would probably be possible to easily resync it on it.
It might be a bit off topic, but why should the BIOS screen be included? The BIOS screen is not part of the game software.
Any runs that use a hard reset sequence would also be "faster" than a real time run, due to not having to play the BIOS. Of course, this run doesn't use it, but I would count personally for consistency to what happens on actual console.
Aren't real time runs always ignoring the BIOS splash, and even the intro, tile screen an loading time? We never include the "host system booting" part in Windows TASes.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
klmz wrote:
It might be a bit off topic, but why should the BIOS screen be included? The BIOS screen is not part of the game software.
The BIOS execution does influence game execution. In fact, some movies would desync if the BIOS were removed. Since we aim for the highest emulation accuracy possible, including the BIOS for all TASes is highly encouraged. By the way, Windows is a special case, so it really doesn't apply here.
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Post subject: Movie label
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
I just figured this movie should get an appropriate label, since it's skipping more than half of the route required for Ending 3... In my opinion "game end glitch" fits nicely. Any thoughts? Edit: According to the Publisher Guidelines, if the glitch used belongs to a common glitch type, that glitch type name should be used. In other words, "game end glitch" is preferred over more the more specific "text glitch" or "hut glitch".
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Dwedit
He/Him
Joined: 3/24/2006
Posts: 692
Location: Chicago
Very Nice!
Post subject: Re: Movie label
fsvgm777
She/Her
Senior Publisher, Player (226)
Joined: 5/28/2009
Posts: 1213
Location: Luxembourg
ThunderAxe31 wrote:
I just figured this movie should get an appropriate label, since it's skipping more than half of the route required for Ending 3... In my opinion "game end glitch" fits nicely. Any thoughts?
Then from my understanding, it no longer obsoletes klmz's TAS, since it performs the Ending 1 route without the "hut glitch" (which doesn't work in the USA release anyway). (and I'd have to re-do the encodes again...)
Steam Community page - Bluesky profile Oh, I'm just a concerned observer.
Patashu
He/Him
Joined: 10/2/2005
Posts: 4043
My thoughts: It's OK for this run to obsolete klmz's one. If an 'all endings' TAS gets made later, it will include gameplay similar to the klmz TAS, since the hut glitch only lets you get endings 1, 3 and 7 and doesn't help you progress in any other endings.
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
Post subject: Re: Movie label
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
fsvgm777 wrote:
ThunderAxe31 wrote:
I just figured this movie should get an appropriate label, since it's skipping more than half of the route required for Ending 3... In my opinion "game end glitch" fits nicely. Any thoughts?
Then from my understanding, it no longer obsoletes klmz's TAS, since it performs the Ending 1 route without the "hut glitch" (which doesn't work in the USA release anyway).
No, it still obsoletes klmz's TAS, because the main goal is beating the game as fast as possible, not doing it in a certain way. Also, we're in Vault territory, for which only two branches are allowed unless the game features extra modes, which is completely unrelated to the ending possibilities. So only fastest completion and best ending are allowed for Vault publications for this game, and this run still counts as Vault fastest completion.
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Post subject: Re: Movie label
Editor, Emulator Coder, Expert player (2156)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
ThunderAxe31 wrote:
No, it still obsoletes klmz's TAS, because the main goal is beating the game as fast as possible, not doing it in a certain way. Also, we're in Vault territory, for which only two branches are allowed unless the game features extra modes, which is completely unrelated to the ending possibilities. So only fastest completion and best ending are allowed for Vault publications for this game, and this run still counts as Vault fastest completion.
Hold on. Does it mean that we can accept "bad endings" aiming for the Vault publications for games that currently only have "good endings" aiming for Moon publications? EDIT: Examples are most Metroidvania Castlevania games.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Dwedit
He/Him
Joined: 3/24/2006
Posts: 692
Location: Chicago
So does this mean a glitchless run is not allowed in Vault?
Post subject: Re: Movie label
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
klmz wrote:
Hold on. Does it mean that we can accept "bad endings" aiming for the Vault publications for games that currently only have "good endings" aiming for Moon publications?
Yes, bad ending runs can be accepted as a separate branch from "best ending" movies, assuming the bad ending is achieved faster than the best ending.
Dwedit wrote:
So does this mean a glitchless run is not allowed in Vault?
Correct. Vault does not allow goals that consist in forgoing use of known glitches or tricks. The Vault rule is very simple:
Vault wrote:
  • Goal choice is limited to fastest completion time (any%), or full-completion (such as 100% or best ending)
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
Post subject: Re: Movie label
DrD2k9
He/Him
Editor, Judge, Expert player (2213)
Joined: 8/21/2016
Posts: 1090
Location: US
ThunderAxe31 wrote:
Dwedit wrote:
So does this mean a glitchless run is not allowed in Vault?
Correct. Vault does not allow goals that consist in forgoing use of known glitches or tricks. The Vault rule is very simple:
Vault wrote:
  • Goal choice is limited to fastest completion time (any%), or full-completion (such as 100% or best ending)
Wait...doesn't this mean that multiple goal choices are valid for vault runs? Why can't there be both a glitched and non-glitched run in vault? EDIT: Unless you are claiming a glitched any% version is better than a non-glitched any%....that I agree with. But I don't agree that a glitched any% should obsolete a non-glitched 100% or full completion run.
Post subject: Re: Movie label
Editor, Reviewer, Skilled player (1358)
Joined: 9/12/2016
Posts: 1646
Location: Italy
DrD2k9 wrote:
Wait...doesn't this mean that multiple goal choices are valid for vault runs? Why can't there be both a glitched and non-glitched run in vault? EDIT: Unless you are claiming a glitched any% version is better than a non-glitched any%....that I agree with.
We only accept the fastest run for Vault. It's not about using glitches or not, it's about being faster. However, note that for full completion we have a rule that forbids glitching the intended completion progression: see below.
DrD2k9 wrote:
But I don't agree that a glitched any% should obsolete a non-glitched 100% or full completion run.
Indeed, a run that goes with "game end glitch, best ending" is not allowed for Vault, as we recently clarified in the rules text, explaining that for full completion it is not allowed to use ACE exploits or memory corruption in order to access the relative completion flags directly:
Vault wrote:
    • Full completion criteria must be reached through in-game actions.
      • Arbitrary code execution (ACE) is not allowed for any full-completion category for the Vault. Arbitrary code execution of ROM data, and memory corruption tricks to write to arbitrary memory are also not allowed.
      • If a game has a progress counter, it must be filled by actions that increment the counter in gameplay. Modifying the progress counter through ACE or memory corruption is not allowed.
      • If full completion is counted by obtaining a set of items, these items must be collected through in-game methods. Using memory corruption to place items into an inventory is not counted towards full completion.
      • If full completion is counted by fulfilling a set of flags, these flags must be set through in-game mechanisms. Using memory corruption to set completion state flags is not allowed.
      • If a progress counter is filled by collecting a set of items or fulfilling a set of flags, all individual components of this set must be collected or fulfilled. Collecting or fulfilling the same component multiple times to inflate the progress counter is not counted towards full completion.
However, glitches that don't modify the relative progress value or completion flags are allowed. As an example, see this recently accepted submission: #5910: luckytyphlosion's GB Pokémon: Blue Version "Gotta Catch 'Em All!" in 37:55.33. It makes use of heavy glitch abuse, however none of these are touching the flags relative to game completion.
my personal page - my YouTube channel - my GitHub - my Discord: thunderaxe31 <Masterjun> if you look at the "NES" in a weird angle, it actually clearly says "GBA"
DrD2k9
He/Him
Editor, Judge, Expert player (2213)
Joined: 8/21/2016
Posts: 1090
Location: US
Makes sense. I was primarily emphasizing that I feel an any % run and a 100% run should both be vault eligible. I don't mind if glitches are used to accomplish the 100% so long as all necessary completion criteria to meet a true 100% run are performed (glitch or not).
Post subject: Re: Movie label
Skilled player (1176)
Joined: 5/11/2011
Posts: 427
Location: China
ThunderAxe31 wrote:
We only accept the fastest run for Vault. It's not about using glitches or not, it's about being faster. Yes, bad ending runs can be accepted as a separate branch from "best ending" movies, assuming the bad ending is achieved faster than the best ending.
I know some games that you can lose everything and still can touch the end, like some sports game. Could we allow only accept the fastest? If we have a tas for moon, could we can accept a bad ending? No matter what it's for Vault or not?