Final Fantasy III (DS)

2015-02-05 02:36:39 am

Speedrunning Final Fantasy III DS

It is pretty much always smarter to fight the enemies, the reason is that (especially early game up to Doga's Grotto maybe/Ancient's Maze) you'll usually 1-shot the enemies. It takes around 3 run aways, before you manage to do so, in that time, most enemies will have time to attack you. But if you can use your first, second (and sometimes third) attempts to kill one enemy after the other, the chances of them even attacking you becomes MUCH smaller. This way, you don't have to open up your menu and heal so often. The times you should run away are when there is overworld encounters, a lot of these is just easy to just run away from. Also usually on preemptive strikes, you should run away, if less it is an enemy that gives A LOT of experience. I usually run away from 1/6 battles throughout the game.

Also, the grinding against the kum kum's are not really necessary, imo.

soloman

2015-02-05 04:46:28 am

i got a few people into running it but only for a certain category. i'd like to open a race category named water crystal %. its basically up to kraken. this category is a bit like the "kefka at narshe" category for FF6

Desch

2015-02-05 05:01:37 am

Speedrunning Final Fantasy III DS

for that race, is the time supposed to end at the Water Crystal? Like, do we need to prepare for late game at all, or only focus on getting to the Water Crystal as fast as possible?

WasabiKnight

2015-02-05 06:12:50 am

I'm totally up for the race, Kraken shouldn't take much longer than 2 hours so it'll be interesting to do Cheesy

I guess if the race is only up until Kraken then we shouldn't need to prepare for late game and just go as fast as possible to Kraken, right?

Also I've found that we probably don't need to pick Cure (or Cura, one or the other since having 1 heal spell is fine) ever. I never need to use it except right before the Giant Rat fight and Medusa, so either picking 1 Cure or 1 Cura should be enough, but I'll try this is out on my next run today.

One last thing, if it's not really necessary to grind Kum Kum's shouldn't we update the document accordingly? I was also extremely confused with the instructions for the last 4 boss fights + CoD, I died the first time because I didn't understand that in the first two fights we were going with 4 SCH + 1 Knight. T_T

soloman

2015-02-05 08:56:22 pm

haha. make your own strats for this. but this will be its own catagory for people not interested in running the full 8 hours of gameplay :p and as a race its 100% viable. unless you guys have any objections to this catagory. you guys can decide

soloman

2015-02-07 05:21:17 am

i managed to recreate a 2 encounter at heins castle. im close to figuring out a way to cause manipulation

Desch

2015-02-07 10:17:09 am

Speedrunning Final Fantasy III DS

Looking forward to seeing the results of your research, Soloman.

But about the run itself, I think that an ultimate time for a SS run of this game would be at around 6:50, assuming no deaths, no screw ups and not too many tricky situations late game. Using Wasabii Knight's latest run as a reference, he used 2 hours and 30 mins from the start of Crystal Tower (that would be AFTER Eureka), to finish. My personal new best time to get to this point is 4:21, so adding that to WasabiKnight's rest makes it at ca 6:50. Another guy at the other SDA thread of this said he had a final save outside of Crystal Tower at 4:27, and even that he thought he could still improve it by 15-30 minutes. I have no idea how he could've done it so fast and still save more time, especially if he didn't know about the Natuilus Dive glitch at that point. He may have skipped ToT, Unei's Cave and Ancient's Ruins to get a time like that, but then the time would be much worse, because of the lack of good armor that is available at that point in the game.

Anyways, about those races, like beating Kraken, I propose another race - A race to beat Titan! When your Natuilus can dive, you actually get access to his boss throught the dive glitch, and I think it would be interesting to see how fast we could do that. I'm estimating 3:20 or maybe even lower SS. Doing the full run is very tiresome (especially in 1 go), so having shorter runs like this could be fun when waiting for other things to happen. What do you guys think?

Looking forward to seeing the results of your research, Soloman.

But about the run itself, I think that an ultimate time for a SS run of this game would be at around 6:50, assuming no deaths, no screw ups and not too many tricky situations late game. Using Wasabii Knight's latest run as a reference, he used 2 hours and 30 mins from the start of Crystal Tower (that would be AFTER Eureka), to finish. My personal new best time to get to this point is 4:21, so adding that to WasabiKnight's rest makes it at ca 6:50. Another guy at the other SDA thread of this said he had a final save outside of Crystal Tower at 4:27, and even that he thought he could still improve it by 15-30 minutes. I have no idea how he could've done it so fast and still save more time, especially if he didn't know about the Natuilus Dive glitch at that point. He may have skipped ToT, Unei's Cave and Ancient's Ruins to get a time like that, but then the time would be much worse, because of the lack of good armor that is available at that point in the game.

Anyways, about those races, like beating Kraken, I propose another race - A race to beat Titan! When your Natuilus can dive, you actually get access to his boss throught the dive glitch, and I think it would be interesting to see how fast we could do that. I'm estimating 3:20 or maybe even lower SS. Doing the full run is very tiresome (especially in 1 go), so having shorter runs like this could be fun when waiting for other things to happen. What do you guys think?

soloman

Edit history:

soloman: 2015-02-08 02:21:39 am

soloman: 2015-02-08 02:21:31 am

2015-02-08 02:18:16 am

the fastest time for completing this game was done by eltherian i beileve at the time 7:50:12. also about the titan race i dont think thats a great idea due to 2 painful factors that make that part of the game lame. and that's temple of time and garuda. up to kraken is good because the only things that can end you is basically 5% hein blizzara, kraken himself, or back attacks when you are going to ancient village. so its more execution based up to that point with not that much death luck as a death puts you behind. though doing a run up to titan is a idea to be tampered with. only way really is to try it and see how it goes. so far the only catagorys right now to my knowledge is water cyrstal, full run, and crystal tower.

Desch

2015-02-08 04:39:52 am

Speedrunning Final Fantasy III DS

Actually, WasabiKnight finished this game at 7:45:25 this friday, which is a better time. And I was only estimating the ultimate time SS, which probably is sub 7 hours.

But that aside, I think just playing up to Kraken is kinda a lame idea. The only real reason is that there's no real challenge to it. Barely any RNG to screw you over, which means that you can easily get the record on it with minimal dedication. It also kinda dull, just to stop playing halfway through the game, I don't think I will ever get a good feeling by stopping at Kraken instead of keep on going. Whereas to beat Titan, you'll be blessed by all 4 crystals, be right beside the final dungeon, and you get that feeling of almost being done with the game. You also have some challenging boss fights that makes it harder to actually get the best time, like Kraken, Garuda and lastly Titan (due to your low levels). Also, you don't have to do the ToT in the Titan run, because you can just go straight to Titan right after the Cave of the Circle, because you can glitch your way into the Ancient's Maze (and Crystal Tower, or both at the same time for that matter). So the Titan run is not as bad as you'd think it'd be.

Actually, WasabiKnight finished this game at 7:45:25 this friday, which is a better time. And I was only estimating the ultimate time SS, which probably is sub 7 hours.

But that aside, I think just playing up to Kraken is kinda a lame idea. The only real reason is that there's no real challenge to it. Barely any RNG to screw you over, which means that you can easily get the record on it with minimal dedication. It also kinda dull, just to stop playing halfway through the game, I don't think I will ever get a good feeling by stopping at Kraken instead of keep on going. Whereas to beat Titan, you'll be blessed by all 4 crystals, be right beside the final dungeon, and you get that feeling of almost being done with the game. You also have some challenging boss fights that makes it harder to actually get the best time, like Kraken, Garuda and lastly Titan (due to your low levels). Also, you don't have to do the ToT in the Titan run, because you can just go straight to Titan right after the Cave of the Circle, because you can glitch your way into the Ancient's Maze (and Crystal Tower, or both at the same time for that matter). So the Titan run is not as bad as you'd think it'd be.

Desch

Edit history:

Desch: 2015-02-08 03:06:16 pm

2015-02-08 07:09:47 am

Speedrunning Final Fantasy III DS

I have now finished my first run to beat Kraken, just to prove how little of a challenge it is. Granted you were supposed to stop the time at the first flash after you beat Kraken, I ended up with a time of 2:10:30. This was my first run were I only duped a few items and skipped some, like Lamia Scale. It's not bad for being the first time, but I got way too many encounters in the last cave, and I even ended up having to do the Kraken fight with 3 characters, since Ingus got petrified, but I still won, luckily. So I guess this is a WR for now and one I'm actually satisfied with, as the Hein's Castle was done with only 3 encounters.

I also noticed 2 wrong things in the Wiki for this game; Djinn's uncommon drop is not Bomb Arm, it is Sheep Pillow, and Giant's Rat can use Fira which overkilled one of my characters at the time.

And here's the "proof" of me winning at that time (in the attached file), but since I don't have a capture card, I can only provide pictures, so I hope that it is enough evidence.

I have now finished my first run to beat Kraken, just to prove how little of a challenge it is. Granted you were supposed to stop the time at the first flash after you beat Kraken, I ended up with a time of 2:10:30. This was my first run were I only duped a few items and skipped some, like Lamia Scale. It's not bad for being the first time, but I got way too many encounters in the last cave, and I even ended up having to do the Kraken fight with 3 characters, since Ingus got petrified, but I still won, luckily. So I guess this is a WR for now and one I'm actually satisfied with, as the Hein's Castle was done with only 3 encounters.

I also noticed 2 wrong things in the Wiki for this game; Djinn's uncommon drop is not Bomb Arm, it is Sheep Pillow, and Giant's Rat can use Fira which overkilled one of my characters at the time.

And here's the "proof" of me winning at that time (in the attached file), but since I don't have a capture card, I can only provide pictures, so I hope that it is enough evidence.

Attachment:

Kraken WR.docx

WasabiKnight

2015-02-08 07:14:38 am

We can always try both races, I mean going to Titan is just one more hour over Kraken, but I also think that it's a pain to go so far because of how bad RNG we can get on Kraken and Garuda so, in my opinion, we should try to race first to Titan and if we end up feeling like it's too exhausting we can change to racing up to Kraken.
Because, once again in my opinion, the fun of racing up to Kraken is that it's a more casual race other than actually competing for the whole game and committing so many hours in to a run we're not actually going to finish. Also I want to note that, by having a Kraken% category we can ease new comers in to the game since you won't have to spend over 8 hours running the game if you make a lot of mistakes at the start.
But I really don't know why you guys want to race so...

TL;DR Kraken% is good for small, casual races (helps people who are learning the game) and Titan% is more competitive and should be more intense to people who know the route well.

I forgot to mention before, I asked Vulajin for control over the first post and I'm updating it with useful information for the game so we don't have everything dispersed in replies.

One last thing, like Desch mentioned, my race was around 7:45 and I'm going to put it on the first post so anyone can use it as reference. You can watch the run at my stream on this highlight right here.

We can always try both races, I mean going to Titan is just one more hour over Kraken, but I also think that it's a pain to go so far because of how bad RNG we can get on Kraken and Garuda so, in my opinion, we should try to race first to Titan and if we end up feeling like it's too exhausting we can change to racing up to Kraken.
Because, once again in my opinion, the fun of racing up to Kraken is that it's a more casual race other than actually competing for the whole game and committing so many hours in to a run we're not actually going to finish. Also I want to note that, by having a Kraken% category we can ease new comers in to the game since you won't have to spend over 8 hours running the game if you make a lot of mistakes at the start. 
But I really don't know [b]why[/b] you guys want to race so...

TL;DR Kraken% is good for small, casual races (helps people who are learning the game) and Titan% is more competitive and should be more intense to people who know the route well.

I forgot to mention before, I asked Vulajin for control over the first post and I'm updating it with useful information for the game so we don't have everything dispersed in replies.

One last thing, like Desch mentioned, my race was around 7:45 and I'm going to put it on the first post so anyone can use it as reference. You can watch the run at my stream on [url=http://www.twitch.tv/wasabiknight/c/6061410]this highlight right here.[/url]

soloman

2015-02-08 07:10:13 pm

as i stated nothing is bad until it is tested. i am willing to try run titan. also nice time wasabi :D. no idea is truly terrible until it is actually performed with a bad result. aside from this. i MANIPULATED GARUDA. http://www.twitch.tv/soloman231/c/6074668
screwed up the first fight cuz Phoenix down. i swear if this does not work for you guys ima be sad.... very sad. oh and i was very happy in the video.

WasabiKnight

2015-02-09 10:31:18 am

Great job on this soloman Cheesy

I will be streaming some more runs this week so I can test incorporating both the glitch in and manipulating Garuda so I'll definitely try out what you did right here Cheesy

the nightmare is finally over...

Eltherian

2015-02-09 10:27:22 pm

Velocity 2X

Hey everyone! I just wanted to pop in and congratulate you guys on the progress made so far! I know that I've kinda fallen off the radar (again), but as sort of an update I lost interest in FF3DS due to time constraints in my life. I don't really have the time for such a long run of any game, but I am still satisfied with my contribution to the game and checking on this thread from time to time to see the progress. I hope that my route has served you guys well (even if you found better optimizations) and that you have a break-through on RNG manipulation and Nautilus Dive soon. Be sure to pass me your Twitch channels so I can give you guys a follow and hop in from time to time to see how things are going. Again, you guys should be proud of the progress that has been made over the past few months and keep it up! I know you can get this thing under 7 hours easy.

WasabiKnight

2015-02-10 07:11:06 am

Hey Eltherian!

It's a pleasure to see you around! I am very thankful to you and Desch contributions to the planning/routing process of this game. The main route is pretty much the same so if you know the old one you know this one as well, because all we've been doing is optimization things here and there, like changing early to a White Mage, or using 4 Scholar's against Kraken for a more consistent fight. The route path is still the same, because we believe this one is the best, so we've been mostly looking for ways to speed up things as much as possible! The one big thing we're changing is we will try to avoid Ancient's Maze altogether, and we don't grind at Crystal Tower anymore, also soloman just discovered how to manipulate Garuda so that removes the big RNG fight off the game entirely!

My stream is http://twitch.tv/wasabiknight I *should* be streaming twice a week so I'll be happy to see you stop by Cheesy

Desch

2015-02-10 10:43:07 am

Speedrunning Final Fantasy III DS

Nice seeing you "active" again, Eltherian, even though you probably won't be much after this. You contributed a lot, especially with the document. The guide is still a WIP, but I think it is soon finished. We have like many different routes as well, which is cool. Kraken%, Titan%, any% and the route that no one ever will do: Iron Giant%. I think that I will one day do the last one, but not in the near future, that's for sure Tongue

Anyways, about that RNG with Garuda, it didn't work at all, lol. For RTA it is viable trying stuff like that, but I believe the game works very different from what you've assumed it to be. When you had a save file and hard resetted your game, you always had the same specific RNG cycle that you had at that save, which means what you did only applied to that one run you did there. I doubt you'll find anything useful as controlling the RNG of this game is insanely hard. Both me and WasabiKnight has checked, so sorry, I guess Sad

soloman

2015-02-10 03:50:51 pm

i had a feeling that it wouldn't work. oh well. i aint giving up just yet. anywho so my list of things to do is 1. encounter manipulation 2. garuda manipulation <--- this is possible 3. i think this one will definitely turn the route around i wanna see if its possible to manipulate the viking through inputs/ items useage

FF3_DS

2015-02-10 04:36:21 pm

If you want to get into specifics of RNG manipulation you need to set what people are playing on because the DSoriginal / DSLite and the 3DS(XL) do not play exactly the same. I'm not totally sure on the emulation side of things but I assume the people streaming are using emulators so... What you do there specifically does not transfer over to the consoles, as stated above.

Yes, Garuda is manipulatable. I have known this for a while, as well as forcing RNG drops from random battles. What people need to take note of is the point of RNG manipulation is that you need to have a consistent run from start to finish to be able to do this in the first place.

Some bosses are not RNG consistent (even if the random battles are). I noted Kraken will never do the same thing after each successive attempts, and Djinn does it as well. I'm not sure what causes this.

soloman

2015-02-10 05:10:31 pm

good points. i'm gonna think of ways to manipulate viking provoke. perhaps a strat like joker doom manipulation in ff6 is possible?

Desch

2015-02-16 11:07:25 am

Speedrunning Final Fantasy III DS

Ok, I have found something and it works. I have tested with both my copies of the game (both the USA version) on both my DSi and my 3DS, always giving me the same identical result. This was tested on WasabiKnight's Emulator, but it did not give the same results, but as far as we know, it can be RNG'ed the same way as the DS version, but it gives different results. Onto the manipulation itself.

I have found out that the following factors does not affect your current RNG cycle: taking more/less time, moving your cursor around, changing weapons (in battle), giving your character different names, and probably some other things. What does actually affect it is opening the Menu and more importantly the way you die. I have not done much research on this, but I think most of this is correct. Whenever you start your game, it will always have the same RNG-cycle. Which means if you start a speedrun right after starting the game, the beginning will be the exact same. But if you die before you start, you will have a different RNG-cycle to start off your game with.

So to put this to the test, I want you guys to check if you get the same results as me when doing this. Start by starting your game and press "New Game". Then you proceed to kill yourself in 2 turns, by hitting yourself twice. After you've died, DO NOT RESET YOUR GAME, because this will reset your RNG-cycle back to default, instead you get back to the title screen and once again press "New Game". After this, you should be faster than the goblins and be able to just mash the A-button for a fast win in the first battle. After this, go in a straight line until you are right next to the stone. Do no turn around to press the stone, just walk straight until you can press A on it. After this, just go to the right passageway and BOOM! You should get Back Attacked by one single Goblin.

As I said, this works on the DS version (only tested on the USA version), and it doesn't give the same results on the emulated version. However, the emulator can be RNG'ed the EXACT SAME WAY as the DS version, it is just that it gives you a different outcome. If this works, it means that we can all work on a way to get the most optimal beginning, and maybe even far into the game, with a very optimal time. If all of us work on different ways to kill yourself to set up a good RNG-cycle to start the run, then we may have something super-sturdy to keep working on. Optimally, we'll have a RNG-cycle that lasts the entire game, (which would be slightly boring) but that's another story. I hope you can help here and contribute and get satisfying results with this method. Looking forward to the results!

Ok, I have found something and it works. I have tested with both my copies of the game (both the USA version) on both my DSi and my 3DS, always giving me the same identical result. This was tested on WasabiKnight's Emulator, but it did not give the same results, but as far as we know, it can be RNG'ed the same way as the DS version, but it gives different results. Onto the manipulation itself.

I have found out that the following factors does not affect your current RNG cycle: taking more/less time, moving your cursor around, changing weapons (in battle), giving your character different names, and probably some other things. What does actually affect it is opening the Menu and more importantly the way you die. I have not done much research on this, but I think most of this is correct. Whenever you start your game, it will always have the same RNG-cycle. Which means if you start a speedrun right after starting the game, the beginning will be the exact same. But if you die before you start, you will have a different RNG-cycle to start off your game with.

So to put this to the test, I want you guys to check if you get the same results as me when doing this. Start by starting your game and press "New Game". Then you proceed to kill yourself in 2 turns, by hitting yourself twice. After you've died, DO NOT RESET YOUR GAME, because this will reset your RNG-cycle back to default, instead you get back to the title screen and once again press "New Game". After this, you should be faster than the goblins and be able to just mash the A-button for a fast win in the first battle. After this, go in a straight line until you are right next to the stone. Do no turn around to press the stone, just walk straight until you can press A on it. After this, just go to the right passageway and BOOM! You should get Back Attacked by one single Goblin.

As I said, this works on the DS version (only tested on the USA version), and it doesn't give the same results on the emulated version. However, the emulator can be RNG'ed the EXACT SAME WAY as the DS version, it is just that it gives you a different outcome. If this works, it means that we can all work on a way to get the most optimal beginning, and maybe even far into the game, with a very optimal time. If all of us work on different ways to kill yourself to set up a good RNG-cycle to start the run, then we may have something super-sturdy to keep working on. Optimally, we'll have a RNG-cycle that lasts the entire game, (which would be slightly boring) but that's another story. I hope you can help here and contribute and get satisfying results with this method. Looking forward to the results!

FF3_DS

Edit history:

FF3_DS: 2015-02-17 10:03:43 am

FF3_DS: 2015-02-17 10:00:55 am

2015-02-17 10:00:23 am

The DSOriginal / DSLite are the same. The 3DS(XL) does not emulate exactly the same way, you probably found a path that led to a shared outcome. I've been saying what you are saying in my previous posts, that the general method of manipulation can be applied across all platforms, but not the specifics.

RNG on a hard reset will always be the same provided your starting point is always the same (at a new game or a hard save right before Garuda). I just ran a quick test and on the DS I did not have time to act on the second attack on myself before the goblins killed me. On the 3DS I managed to kill myself on the second attack. After the game over: DS did not yield attacks before goblins on rounds, however, on the 3DS it did.

EDIT: I suggested earlier that it may be that each cartridge itself is unique based on what you stated, as I did attacks on myself on the DS and I didn't act before the goblins.

Desch

2015-02-17 10:24:59 am

Speedrunning Final Fantasy III DS

Are you sure you did the same things as I said? I know that you've said before that you think that each cartridge is unique or have different starting points, but I won't believe that I can RNG manipulate 2 cartridges in the exact same way in that case. Everything that happened on the 2 different cartridges where identical all the way down to damage, turn order, crits, back attacks and so on.

So I'm just gonna ask again, did you do exactly as I suggested? And did you not get the back attack? (<-- easy way to know if it works)

FF3_DS

2015-02-17 10:27:53 am

I already know the outcome is going to be different if the goblin acts out of order. It was not identical on either system.

Desch

2015-03-13 04:35:04 pm

Speedrunning Final Fantasy III DS

Ok, about the RNG, turned out I was right. What made the difference was that we played on different consoles, which gave us different base RNG-cycles.

Anyways, here's an updated version of the guide which should be used from now on. There's a couple of changes, so be sure to read it, and that includes how to RNG manipulate the game.

Attachment:

FF3 DS Speedrun Route rev4.0.docx

Desch

Edit history:

Desch: 2015-03-25 07:51:24 am

Desch: 2015-03-24 09:07:46 am

2015-03-24 09:05:01 am

Speedrunning Final Fantasy III DS

I need your opinion on pre-game/RNG setup! Originally I thought that this should not be counted as total play time, but I have now got some insight from other's and can see this being argued in both ways. So should pre-game RNG setup be counted as the run, or doesn't it count?

Why it should not be timed:
- Opens much more "easy" diversity.
- It does not favour one console much more than any other.
- You don't have to reset the game every time you decide to restart (not a big problem, tho).
- We can keep our simple definition of the run as it is.
- It counts as separate preparations for the run.

Why it should be timed:
- You use time to set up the RNG
- You have already pressed "New Game"
- Doesn't count as separate preparations for the run.

Can't think of more reasons for any of the two at the moment from the top of my head. Anyways, I just want some discussion and opinions on this. Personally I'd prefer it if it would not be timed.

Also, the more I think about when we should stop the timer, the more I think we should time it either after we've saved after the credits, or when you press B when you are given the option to save, but I guess that this can be counted as in-game time instead.

I need your opinion on pre-game/RNG setup! Originally I thought that this should not be counted as total play time, but I have now got some insight from other's and can see this being argued in both ways. So should pre-game RNG setup be counted as the run, or doesn't it count?

Why it should not be timed:
- Opens much more "easy" diversity.
- It does not favour one console much more than any other.
- You don't have to reset the game every time you decide to restart (not a big problem, tho).
- We can keep our simple definition of the run as it is.
- It counts as separate preparations for the run.

Why it should be timed:
- You use time to set up the RNG
- You have already pressed "New Game"
- Doesn't count as separate preparations for the run.

Can't think of more reasons for any of the two at the moment from the top of my head. Anyways, I just want some discussion and opinions on this. Personally I'd prefer it if it would not be timed.

Also, the more I think about when we should stop the timer, the more I think we should time it either after we've saved after the credits, or when you press B when you are given the option to save, but I guess that this can be counted as in-game time instead.