So this is a project I'm looking to pick up and finish at least one category for (probably Any%, 100% is definitely an option though). I was wondering if anyone knew of any sort of notes documentation from previous projects people did that they could post in here. Would save me a ton of time research-wise and let me skip straight to re-route/optimizations. I've already been through this thread and jotted down a handful of things such as the fisherman RNG testing/addresses, the boss strategy developments, and other potentially useful stuff.
That being said though, I don't have any of the critiques Nitrodon made of Dessyreqt's WIP, which knowing Nitrodon, are probably very valuable to have. I don't know anything Dessyreqt might have found that was optimal in the progress they made. Etc...
So yeah, if anyone has anything at all, I would absolutely love to get my hands on it to speed this process up. Going to go scour through the SDA forum thread for this game now to try and find some more fun stuff there.
Thank you in advance.
Update #1, Or something....:
So it seems as though the RTA waits a specified amount of time through the demo that plays from the title screen in order to assure a specific RNG roll when they head towards the cave to spawn the fisherman on RNG call 52.
The TAS, however, starts at power on. I don't plan to wait through the intro because a) it's boring and b) I just don't want to. So that's obviously my stance on it.
However, I am curious what all of you think about this and am entirely open to suggestions on what to do. Actual work on this TAS should start next week sometime unless something comes up.
Joined: 3/2/2010
Posts: 2178
Location: A little to the left of nowhere (Sweden)
A good idea may be to hop on IRC and try to catch dessyreqt, alternatively try to PM him, in case he has some notes you could use.
Though, looking at his last sign-in date, IRC would be a safer bet.
So just curious, out of the people that would be interested in seeing this game have a TAS that's a little more up to date with various things.
Any% or 100%?
I started working on Any%, and haven't really made it too far mostly because I'm not sure Any% is different enough to merit having it's own category on the site. I would really appreciate some feedback on this from the people who want to see it get done. I'd rather start over and go the 100% route now instead of halfway through.
Project is "delayed" until I figure out what the majority response is.
Also, if it does merit separate categories, I already intended on doing both, but doubling my workload just delays future projects.
Thank you in advance.
I don't really think that any% and 100% are well-differentiated in this game. 100% just involves going out of your way at various points to grab red jewels, and then doing an extra dungeon at the end.
That said, any% and 100% are the two categories that are automatically Vault-eligible, so I wouldn't worry about your run not getting published either way.
I don't feel too strongly about which one you do. The extra dungeon isn't super-interesting IMO.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I'm pretty much in agreement with what you've said.
I've been talking with a few people on skype/steam about this "issue" as well, and I think I've just come to the conclusion that I'll just do them both side-by-side since they're really not that different. That way, both get knocked out, and IF anything awesome happens that does differentiate the two, we end up having our 2 categories. And if nothing ends up differentiating them, then we'll have both categories in the vault with updated strategies, routes, and whatnot.
I'm cautiously optimistic about this particular game not taking too terribly long to TAS, but who knows what will come up.
Joined: 12/8/2012
Posts: 706
Location: Missouri, USA
I'm honestly a bit more partial to the 100% (probably because I played this one a lot, back in the day, and strove to collect all of those red jewels). But as Derakon stated, either category is eligible for publication to the Vault (and Moons, based on majority positive voting/feedback).
Panda, I also wanted to go ahead and give a quick TASing tip that I noticed during your SoE run: In the very beginning, I noticed that you had the start key held down in order to get through the intro and title screen. If you instead though, through trial and error hit the start key on the first possible frame that will get through those screens, you'll very likely save time (since that singular input will register quicker than having the key held down, throughout).
Looking forward to your runs :)
"But as it is written, Eye hath not seen, nor ear heard, neither have entered into the heart of man, the things which God hath prepared for them that love him." - 1 Corinthians 2:9
I'm honestly a bit more partial to the 100% (probably because I played this one a lot, back in the day, and strove to collect all of those red jewels). But as Derakon stated, either category is eligible for publication to the Vault (and Moons, based on majority positive voting/feedback).
Panda, I also wanted to go ahead and give a quick TASing tip that I noticed during your SoE run: In the very beginning, I noticed that you had the start key held down in order to get through the intro and title screen. If you instead though, through trial and error hit the start key on the first possible frame that will get through those screens, you'll very likely save time (since that singular input will register quicker than having the key held down, throughout).
Looking forward to your runs :)
I'm going to end up doing both anyways, so I'll just do them side-by-side and the votes can decide. It's really not that much more work for this game since I have a fairly decent grasp of how the RNG works and stuff like that. The exception being if something broken is found and suddenly alters how close the two categories are.
As for the SoE thing, I actually double-checked and ended up having the same frame advance the scene with or without holding Start down. This is something I've noticed though in other areas of that TAS, and have taken note of it and added it to the ever-growing list of things I check out before progressing past that point. Thank you for pointing it out though, much appreciated.
edit1: This is unfortunately delayed for now, having some issues with my computer at the moment. I'm really enjoying TAS'ing this game though, so hopefully I can get back to it asap.
Up!
I have started an Any% TAS of IoG. You can check out the bk2 file in my WIPs. Here's a YouTube encode of the run so far:
https://www.youtube.com/watch?v=v_WAg3yYDPg
This goes up to the end of Edward's Prison so far. Enjoy!
New WIP! This goes up to the first Mystic Statue in the Gold Ship.
https://www.youtube.com/watch?v=9hcu8nqX4ws
I confirm what Dessyreqt found in his WIP: Will is indeed faster than Freedan to kill Castoth (by 325 frames). I also achieve 16s remaining on the timer for the Moon Tribe Challenge. I also found out that ramp tricks on vertical ramps are actually possible but frame-perfect: you need to be facing away from the ramp and initiate the jump attack on the correct frame. Finally, I realized that you could climb ladders much faster by inputting Up or Down 4 frames out of 5 (leave an empty input every 5 frames). This makes the hero move by 2 pixels 4 frames every 5 frames, instead of 4 every 6 frames.
This is looking pretty fantastic! I'm excited to see the whole run. A couple questions, though. how did you do the super fast dash after first talking with Kara in the palace? And is the pause buffering used to skip the level up animation?
This is looking pretty fantastic! I'm excited to see the whole run. A couple questions, though. how did you do the super fast dash after first talking with Kara in the palace? And is the pause buffering used to skip the level up animation?
For the fast dash, you just need to talk to the guard next to Kara's room. He will tell Will to go away and send him with super fast speed. It's just a normal feature of the game!
The pause trick is called an Orb Cancel. Everytime you kill an enemy which releases an orb to clear a passage, you can simply go to the inventory menu and come back. The animation will still be carried out but the effect will already be in place (for instance, the door will be effectively created even if it is not visible yet).
This new Aura Barrier warp has ACE potential that needs to be analyzed.
Specifically, Aura Barrier warping and saving without fixing the glitched state (by opening the inventory for example) causes a save data corruption that's so bad that the main menu crashes after a hard reset when attempting to print the map name from a save file. (There's no known recovery beyond physically wiping the SRAM at this point.)
However, one of the game's normal routine (a generic block memory copy) is executed from RAM, and since the text parser overwrites RAM, there is a possibility that a useful ACE could be produced from that.
Shutout to Raeven0 for providing this information. More details from him:
For e.g. the "Start Journey" menu, the text script starts at $0bf437. An error is introduced when a map number is read, e.g. for Diary1 at $0bf452 C5 06 F7 74 0D which is parsed as db $C5 : dw $F706, $0D74. $C5 is an indexed-include "opcode", first arg is the address of a pointer list (in the same bank as the caller, here $0b), second arg is the address (same bank again) containing the index of a pointer in that list. Earlier $0D74 is set to the saved map number of file 1. If we have e.g. $0D74 = #$00CC, we read the pointer at $0bf706 + 2*$CC which is $fa3b, so the text parser jumps to $0bfa3b 60 A9 A2 80 8C 88 83 CA which are the bytes for "P y r a m i d" followed by the $CA "opcode" which means "return".
So the ACE challenge is: find a value for $0D74 that causes the text parser to jump to a sequence of bytes that, rather than displaying text, happens to do something more interesting.
After working on a really crazy route for the past 2 weeks, I made a 20:11 TAS that abuses sub-frame resets to duplicate Red Jewels:
http://tasvideos.org/userfiles/info/68531266844059737
This is possible because a diary's checksum does not change when swapping bits in the same 16-bit position in SRAM, allowing for save corruption exploits. This checksum property was discovered by Raeven0.
I don't really have the time to optimize it at the frame level for a submission though, so if anyone wants to team up with me to help me on this regard, please let me know!
Full Any% TAS is back on track! I re-did Incan Ruins with the new strats and manipulated some nice fish RNG for the rafting sequence.
https://www.youtube.com/watch?v=bmLo829A84s
It looks like [4540] SNES Illusion of Gaia "save glitch" by SmashManiac in 16:48.69 is improvable as LeHulk discovered that talking to the last soul in Tower of Babel from the top instead of from the right is faster. I may have messed-up some sub-pixel movement as well, as there was an instance in a Dark Space where I needed an extra frame of movement for no apparent reason.
I'm not planning to update this movie in the foreseeable future, so if anybody wants to take a jab at improving it, be my guest!
