Dragon Warrior Run

2015-12-03 06:44:28 am

I can't see why it would be a separate category. Dragon Warrior 3 uses metal slime manipulation, along with other manipulations, as the normal run.

Lhexa

2015-12-03 07:22:09 am

I say don't make a new category. My guideline is that a category should only be split when there are enough people to do runs in both sides of the split, and while that's the case in Japan, it's not so here.

Also, I'm happy to see so many people interested in the game!

Caracarn

2015-12-04 03:34:53 pm

just a quick update, I've mapped out 2-3 alternate paths for different castle NPC movements (oddly, 2 of the NPC movements lead to the same encounter path). All of the paths lead to the exact same metal slime battle, so you do the same thing in battle no matter which path you get. I have also managed to get a 2nd and 3rd metal slime consecutively, using an audio cue for when to clear the text box, however, the window for doing this is quite small, so it's going to take considerable practice to do this consistently. So far I have not been able to find a 4th metal slime that does not kill me immediately. If we can find a reliable 4 metal slime route, we could then save again and find a manipulation strategy for level 7 Axe Knight! The possibilities are endless!

hurfaderfa

2015-12-05 01:53:45 am

Quote from Highspirits:

I can't see why it would be a separate category. Dragon Warrior 3 uses metal slime manipulation, along with other manipulations, as the normal run.

I can not disagree with this strongly enough. To me there is a fundamental different between manipulations that can be discovered and refined working on a console and those that require using an emulator that displays memory values to map out. This manipulation while theoretically possible to map using a console would be so difficult as to be practically impossible. Assuming I am understanding the way the manipulation works correctly this will inevitably result in, at a minimum manipulation including grinding to 7, defeating the axe knight, and retrieving the sword. That makes this new speedrun not just different but practically unrecognizable compared to the current run. It's similar to the difference other games have between glitchless and glitched categories both in terms of theoretical run length and the amount of the game world explored. If the worry is about having enough runners for each category I think we need to talk to some runners and see who is willing to put up times in the new category. That isn't really my concern though, as a member of the DW community as it were I would just like to see the current / RNG routes through the game still exist. If they have to compete for time with "human tas" routes I think that will spell the end of what I(and I'm sure many others) think of when I think about DW speedruns.

As a somewhat separate note while I would not at all be opposed to a separate category for manipulation-free DW3(hell, I'd love for it to happen! I'd probably get to watch more DW3 runs) that's kind of a different story. That's a game where the manipulation is all figured out on console and cartridge. The new DW manipulation is figured out using a TAS-style emulator.

[quote="Highspirits"]I can't see why it would be a separate category.  Dragon Warrior 3 uses metal slime manipulation, along with other manipulations, as the normal run.[/quote]

I can not disagree with this strongly enough. To me there is a fundamental different between manipulations that can be discovered and refined working on a console and those that require using an emulator that displays memory values to map out. This manipulation while theoretically possible to map using a console would be so difficult as to be practically impossible. Assuming I am understanding the way the manipulation works correctly this will inevitably result in, at a minimum manipulation including grinding to 7, defeating the axe knight, and retrieving the sword. That makes this new speedrun not just different but practically unrecognizable compared to the current run. It's similar to the difference other games have between glitchless and glitched categories both in terms of theoretical run length and the amount of the game world explored. If the worry is about having enough runners for each category I think we need to talk to some runners and see who is willing to put up times in the new category. That isn't really my concern though, as a member of the DW community as it were I would just like to see the current / RNG routes through the game still exist. If they have to compete for time with "human tas" routes I think that will spell the end of what I(and I'm sure many others) think of when I think about DW speedruns.

As a somewhat separate note while I would not at all be opposed to a separate category for manipulation-free DW3(hell, I'd love for it to happen! I'd probably get to watch more DW3 runs) that's kind of a different story. That's a game where the manipulation is all figured out on console and cartridge. The new DW manipulation is figured out using a TAS-style emulator.

Crow!

Edit history:

Crow!: 2015-12-05 11:37:30 am

2015-12-05 11:36:55 am

What's that gemma?

CaracarnVi's runs are done on console. I used (and Caracarn is using) an emulator to make mapping out the route more convenient, but the difference is simply a matter of time rather than a matter of plausibility - taking like 20 seconds try each value instead of ~240 seconds. The reason this is so easy is that the RNG in Dragon Warrior is incredibly weak. For example, Lhexa figured out an RNG manipulation to kill Metal Slimes during his segmented runs, without having really meant to do so in the first place.

As for whether RNG manipulation should be allowed for games in general, I think the western speedrun community has made it clear that it's fair game - I know many games which use it in the standard routes and I don't even know of a single one which has a "no manipulation" category. A general statement, I think, is that in the western communities, we define run categories based on what the player does, rather than how or when the player does it.

hurfaderfa

Edit history:

hurfaderfa: 2015-12-05 12:32:31 pm

2015-12-05 12:29:06 pm

Quote from Crow!:

CaracarnVi's runs are done on console. I used (and Caracarn is using) an emulator to make mapping out the route more convenient, but the difference is simply a matter of time rather than a matter of plausibility - taking like 20 seconds try each value instead of ~240 seconds. The reason this is so easy is that the RNG in Dragon Warrior is incredibly weak. For example, Lhexa figured out an RNG manipulation to kill Metal Slimes during his segmented runs, without having really meant to do so in the first place.

If you think that figuring out the exact manipulation for the whole walk down to the metal slime area, battle manipulation and then finding the second and third slimes would only take a short time on a console then uh, I don't know what to tell you. That's just wildly false.

Quote from Crow!:

A general statement, I think, is that in the western communities, we define run categories based on what the player does, rather than how or when the player does it.

I agree, in this case I would define the "human tas" category as the one that requires hours of sitting with an emulator testing different memory values to completely break the game. If someone figured out how to completely follow the tas route I wonder if you would feel the same way? The 2 or 3 metal slimes manipulation is just the tip of the iceberg, I suspect that eventually someone will go much further.

[quote="Crow!"]CaracarnVi's runs are done on console.  I used (and Caracarn is using) an emulator to make mapping out the route more convenient, but the difference is simply a matter of time rather than a matter of plausibility - taking like 20 seconds try each value instead of ~240 seconds.  The reason this is so easy is that the RNG in Dragon Warrior is incredibly weak.  For example, Lhexa figured out an RNG manipulation to kill Metal Slimes during his segmented runs, without having really meant to do so in the first place.[/quote]

If you think that figuring out the exact manipulation for the whole walk down to the metal slime area, battle manipulation and then finding the second and third slimes would only take a short time on a console then uh, I don't know what to tell you. That's just wildly false.

[quote="Crow!"]A general statement, I think, is that in the western communities, we define run categories based on what the player does, rather than how or when the player does it.[/quote]

I agree, in this case I would define the "human tas" category as the one that requires hours of sitting with an emulator testing different memory values to completely break the game. If someone figured out how to completely follow the tas route I wonder if you would feel the same way? The 2 or 3 metal slimes manipulation is just the tip of the iceberg, I suspect that eventually someone will go much further.

welcotar

2015-12-05 12:59:32 pm

Hi friends. An interesting route that we were spitballing last night is
1. manip L6 from two metal slimes, then stay at 230 xp --- resist urge to grind for a while.
2. death warp, key, death warp, gold grind. you will want to buy dragon scale, broad sword and large shield only.
3. after buying broad sword, death warp. spend or death down to exactly 0 gold.
4. now you have 230 XP, 0 gold etc. save + reset to manip again. this will get you a consistent seed at L6.
5. manip your way down to the axe knight and kill him (at L6).
6. death back to rimuldar and continue grinding.

I did a theory tas this morning and it seems possible, but alas I am not skilled enough to do an rta. The hardest part is just finding a battle where axe knight doesn't pre-emptively kill you, but if you can find it reliably, it is just buffering to win.

Here is a video of my test battle:
and here is a lua program, loto router, to help route the battle: http://pastebin.com/Z7bbkggN

I verified that I can get the same hits buffering manually (though e.g. "14/3" meant I had to count to 15/4.)
Hope this is of use for routing!

Hi friends. An interesting route that we were spitballing last night is
1. manip L6 from two metal slimes, then stay at 230 xp --- resist urge to grind for a while.
2. death warp, key, death warp, gold grind. you will want to buy dragon scale, broad sword and large shield only.
3. after buying broad sword, death warp. spend or death down to exactly 0 gold.
4. now you have 230 XP, 0 gold etc. save + reset to manip again. this will get you a consistent seed at L6.
5. manip your way down to the axe knight and kill him (at L6).
6. death back to rimuldar and continue grinding.

I did a theory tas this morning and it seems possible, but alas I am not skilled enough to do an rta. The hardest part is just finding a battle where axe knight doesn't pre-emptively kill you, but if you can find it reliably, it is just buffering to win.

Here is a video of my test battle: https://youtu.be/l_uYpuG8sxw
and here is a lua program, loto router, to help route the battle: http://pastebin.com/Z7bbkggN

I verified that I can get the same hits buffering manually (though e.g. "14/3" meant I had to count to 15/4.)
Hope this is of use for routing!

puwexil

2015-12-05 02:56:07 pm

Professional Second Banana

I've never seen a problem with using emulators and other tools to assist in the routing process, as long as the final result is strats that can be done on console in real-time.

Caracarn

2015-12-05 04:22:17 pm

I noticed that the axe knight battle changes every 16 frames, so hopefully it won't be too difficult to use a music cue to initiate the battle, once we find a battle that works. Hauksness music may not be very good for that though

dunnius

2015-12-05 07:32:07 pm

Intruding N313 and F014

I was thinking about the route too, but that new method of metal slime grinding is awesome. The number of slimes for grinding can easily be extended because you can put in lengthy delays if necessary to manipulate luck. Even if you had to put in a delay of a minute for a slime (it would never come close to that, of course) it would still be faster overall than other methods.

Grinding to level 7 for the axe knight is better because of being able to sleep lock, making manipulation easier and a faster fight. But thinking about when to stop grinding metal slimes, I can't tell at what point it is faster to keep grinding slimes for fast exp and get additional gold from the gold glitch, or stop at a lower level and grind normal monsters for exp and gold.

If the route is metal slime grind to lvl 7, then axe knight, and then gold glitch (it should also be possible to manipulate to avoid encounters using audio cues), would getting the large shield be necessary because of the extra defense and healing of Erdrick's Armor? Also, if the metal slime grind went up to level 13, would that remove the need for the broad sword as well?

But the big question is whether the game can now be completed at level 7 if audio cues can be used to eliminate encounters to the Dragon Lord, as well as sleep locking both forms.

I was thinking about the route too, but that new method of metal slime grinding is awesome.  The number of slimes for grinding can easily be extended because you can put in lengthy delays if necessary to manipulate luck.  Even if you had to put in a delay of a minute for a slime (it would never come close to that, of course) it would still be faster overall than other methods.

Grinding to level 7 for the axe knight is better because of being able to sleep lock, making manipulation easier and a faster fight.  But thinking about when to stop grinding metal slimes, I can't tell at what point it is faster to keep grinding slimes for fast exp and get additional gold from the gold glitch, or stop at a lower level and grind normal monsters for exp and gold.

If the route is metal slime grind to lvl 7, then axe knight, and then gold glitch (it should also be possible to manipulate to avoid encounters using audio cues), would getting the large shield be necessary because of the extra defense and healing of Erdrick's Armor?  Also, if the metal slime grind went up to level 13, would that remove the need for the broad sword as well?

But the big question is whether the game can now be completed at level 7 if audio cues can be used to eliminate encounters to the Dragon Lord, as well as sleep locking both forms.

Crow!

Edit history:

Crow!: 2015-12-06 11:56:31 am

Crow!: 2015-12-06 11:52:36 am

Crow!: 2015-12-06 11:52:29 am

2015-12-06 11:51:06 am

What's that gemma?

Unfortunately, the Sleep spell and the Stairs action are both not reliable for RNG manipulation because you must be frame perfect during the menus, which is incredibly difficult. So, messing around in Charlock at comically low levels is, for now at least, off limits.

I think that, rather than including a 16-frame-window wait for a few seconds before the Axe Knight fight, instead repetitive menu accesses via buffering A and B could advance the RNG in a more controlled fashion, which would consistently find some particular Axe Knight fight, which you could then route out the proper "Command?" sequence for.

Also, since you can only move on a 16 (or whatever it is) frame rule, most of the time if you start at a different RNG in the Castle, you won't be able to find the exact same Axe Knight fight. You'll probably need to route out a different "Command?" sequence to defeat the Axe Knight for each of the different initial NPC movements that you map out.

welcotar

2015-12-06 07:45:09 pm

At kirkq's suggestion I made a spreadsheet of L6 axe knight battles: https://goo.gl/f5qNL3

You can look up the rng seed while your battle is loading and then read off the buffer counts for good seeds (about 1/8). I spot checked like 30 or 40 of these, but there might be errors because the program (http://pastebin.com/6mSZh6Fu) simulates battles (emu was too slow). Interestingly it turns out being slept is part of some of the faster battles, so there are some where you crit, fall asleep, tank a hit, wake up, and then crit again.

I am still working on computing routes for the better seeds.

Caracarn

2015-12-06 08:17:11 pm

so far I have a 28/30 on one path, and the same 34/28 on two different paths. Not the greatest Axe Knight battles, but it's a start.

dunnius

2015-12-07 03:58:56 pm

Intruding N313 and F014

But couldn't you wait in the menu with the cursor on the sleep spell for an audio cue?

Caracarn

Edit history:

Caracarn: 2015-12-08 12:07:19 am

2015-12-08 12:03:59 am

So here's my plan for after Level 6 Axe Knight:

Go to Brecconary and buy Dragon Scale, use Dragon Scale, Sell Dragon Scale, Buy 2 herbs
This is a bit out of the way at this point, but we need the Dragon Scale to fight Wolflords before level 10 and Wyverns before Level 12 so I think it's worth it. The herbs are for safety since we don't want to have to stop at an inn during the grinding at all.

Grind to Level 9 outside Rimuldar, go buy 5 keys once you hit level 9. Wolflords are feasible at level 8, but Wraiths at level 8 will be terrible.

Grind to Level 11 in the Wolflord, Wraith, Metal Scorpion zone. At level 9, might have to dip back into the other zone when HP is low.

Grind to Level 13 in the hills around Rainbow Drop Cave, with Wyverns, Wolflords, and Wraiths. Again, might have to go to the weaker zone if HP gets low.

Gold is actually not an issue at all, so no Goldmen need to be killed, we should have well over 15k by the time we hit level 13. Make sure you have kept at least 8 MP to cast return, or 16 to return from Stones of Sunlight room also.

Go get the Stones of Sunlight immediately after returning.

Go to Brecconary and use the inn, and buy 4 Fairy Water

Use Fairy Water and go to Garinham. Buy 6 Herbs, then go get Silver Harp, cast Outside/Return

Use Fairy Water, run to Kol, get Fairy Flute

Use Fairy Water, get Staff of Rain, cast Return

Start running to the Cantlin/Token area, use the last fairy water partway there.
I expressed concern before about getting the Token before Cantlin, cause we risk dying and losing our money, but now I think we should actually get the Token first since we aren't spending MP and Herbs on Axe Knight anymore. Also, we will have more MP for the sword rush without having to restore MP an extra time.

After getting the token, go to Cantlin and kill the Golem. Go to the left first and restock on herbs if needed and stop at the Inn. Then buy the Silver Shield, then buy 4 Fairy Water, Use a Fairy Water, then buy 1 more Fairy Water, then cast Return.

Run to Swamp Cave, use a Fairy Water before it runs out. Run to Rainbow Drop, use a Fairy Water before it runs out. Run to Charlock, get the sword, cast Outside/Return.

Either stop at the inn or use the MP restore in the castle. I think the inn is 2 seconds faster, but only if you don't get an extra encounter due to the longer run, so this is debatable.

Use a Fairy Water and run to the Hauksness desert.

Grind to Level 18. Methods of grinding here are still up for debate. Keep 10 MP.

Cast Return, restock herbs and buy a torch. Run to the Inn, cast Repel, then use the inn.
Use a Fairy water before Repel wears off to make it the rest of the way to Charlock.

Get to Dragonlord and kill him.

When I am done editing my excel chart with all these notes plus maps for the manipulation paths, I will attach it to a post.

So here's my plan for after Level 6 Axe Knight:

Go to Brecconary and buy Dragon Scale, use Dragon Scale, Sell Dragon Scale, Buy 2 herbs
This is a bit out of the way at this point, but we need the Dragon Scale to fight Wolflords before level 10 and Wyverns before Level 12 so I think it's worth it. The herbs are for safety since we don't want to have to stop at an inn during the grinding at all.

Grind to Level 9 outside Rimuldar, go buy 5 keys once you hit level 9. Wolflords are feasible at level 8, but Wraiths at level 8 will be terrible.

Grind to Level 11 in the Wolflord, Wraith, Metal Scorpion zone. At level 9, might have to dip back into the other zone when HP is low.

Grind to Level 13 in the hills around Rainbow Drop Cave, with Wyverns, Wolflords, and Wraiths. Again, might have to go to the weaker zone if HP gets low.

Gold is actually not an issue at all, so no Goldmen need to be killed, we should have well over 15k by the time we hit level 13. Make sure you have kept at least 8 MP to cast return, or 16 to return from Stones of Sunlight room also.

Go get the Stones of Sunlight immediately after returning.

Go to Brecconary and use the inn, and buy 4 Fairy Water

Use Fairy Water and go to Garinham. Buy 6 Herbs, then go get Silver Harp, cast Outside/Return

Use Fairy Water, run to Kol, get Fairy Flute

Use Fairy Water, get Staff of Rain, cast Return

Start running to the Cantlin/Token area, use the last fairy water partway there.
I expressed concern before about getting the Token before Cantlin, cause we risk dying and losing our money, but now I think we should actually get the Token first since we aren't spending MP and Herbs on Axe Knight anymore. Also, we will have more MP for the sword rush without having to restore MP an extra time.

After getting the token, go to Cantlin and kill the Golem. Go to the left first and restock on herbs if needed and stop at the Inn. Then buy the Silver Shield, then buy 4 Fairy Water, Use a Fairy Water, then buy 1 more Fairy Water, then cast Return.

Run to Swamp Cave, use a Fairy Water before it runs out. Run to Rainbow Drop, use a Fairy Water before it runs out. Run to Charlock, get the sword, cast Outside/Return.

Either stop at the inn or use the MP restore in the castle. I think the inn is 2 seconds faster, but only if you don't get an extra encounter due to the longer run, so this is debatable.

Use a Fairy Water and run to the Hauksness desert.

Grind to Level 18. Methods of grinding here are still up for debate. Keep 10 MP.

Cast Return, restock herbs and buy a torch. Run to the Inn, cast Repel, then use the inn.
Use a Fairy water before Repel wears off to make it the rest of the way to Charlock.

Get to Dragonlord and kill him.

When I am done editing my excel chart with all these notes plus maps for the manipulation paths, I will attach it to a post.

dunnius

2015-12-08 04:44:17 am

Intruding N313 and F014

I'm confused, are you saying that the grind to 13 should be done without a weapon?

Caracarn

2015-12-08 06:20:48 am

we have the Broad Sword to kill Axe Knight at level 6 with 2 excellent moves

Crow!

Edit history:

Crow!: 2015-12-08 10:27:38 am

2015-12-08 10:27:14 am

What's that gemma?

Caracarn just failed to list the stuff done up through the Axe Knight, which is:

Collect the chests in the King's chamber, go down then back up the stairs, save, reset

Defeat 3 Metal Slimes via RNG Manipulation, then die (you could also route for 2 kills for fewer headaches at getting a run off the ground, but it requires a completely different Axe Knight manipulation later)

Run to Rimuldar, buy a Key, deathwarp

Set up and do the chest glitch to get enough money to buy all the rest of the stuff you need

Collect the Large Shield, Fairy Flute, and Broad Sword (IMO the Dragon Scale should be purchased here, but that's not what Caracarn is doing at present... again, putting it in will change the Axe Knight manipulation)

Manipulate GP to some consistent value (0 is a sensible choice but not mandatory), then save and reset

Defeat the Axe Knight via RNG Manipulation, then deathwarp

gyre

2015-12-08 01:26:58 pm

Are there viable seeds for killing the Axe Knight with the Hand Axe? Idea would be to do less chest grinding, buy both Axe and Shield from Garinham which is less walking, and then buy the Broad Sword early in the grind at 8 before the keys at Rimuldar.

Caracarn

Edit history:

Caracarn: 2015-12-09 08:43:46 pm

Caracarn: 2015-12-08 05:08:34 pm

Caracarn: 2015-12-08 05:07:59 pm

2015-12-08 05:01:55 pm

max crit damage with hand axe would be 31, so it's definitely still possible, it's just a matter of finding the right setup. There are so many possibilities that it will be hard to find the best option.

Also, I'm not getting the fairy flute before Axe Knight, I'm getting it at level 13 on the way to the Staff of Rain.

As far as the Dragon Scale goes, you don't really need it for Axe Knight, ~~and since you can't buy it in Garinham,~~ (edit: someone told me this for some reason, but you actually can buy it in Garinham) ~~it's just as much out of the way after Axe Knight as it is before.~~ Adding Dragon scale to my current setup gave me really terrible paths ~~also~~, so that's why I took it out.

This is certainly not the final route, but it's a good start at least. I just don't know how much I want to keep grinding out all the possible seeds. If someone else wants to go for it, I'm all for it though thumbsup

Caracarn

2015-12-09 08:41:11 pm

Completed a run with the new manipulation strats: http://www.twitch.tv/caracarnvi/v/29148296 - 4:57:08

This was with the epic blooper of forgetting to grab the armor before walking off the tile, and having to do the Axe Knight again. There was also a death at level 14 and level 17, and 3rd try Dragonlord (1st try died to DL, 2nd try died to Armored Knight on last floor).

welcotar

2015-12-09 09:12:28 pm

!! congrats Caracarn. I can't wait to see where this goes next. Maybe it'll even hit 4:30...

When I am not busy pretending to be a productive member of society, I am working on some programs to help with routing more paths. Hand axe is sounding like a really good idea... :-)

Lhexa

2015-12-10 09:33:05 pm

The Dragon Warrior speedrunning community is now large enough to have serious discussions about category definitions! *beams with pride*

To any segmented runners, I do request that you still try to implement my weird, accidental Metal Slime manipulation in the Hauksness desert grind, since I consider that my sole contribution to the speedrunning lore.

Caracarn

2015-12-15 05:14:14 pm

Here's my (hopefully) finished Dragon Warrior RTA Chart, with notes and maps. Obviously, this is subject to change if someone finds better Metal Slime/Axe Knight paths, or adds other manipulation into the route.

Attachment:

Dragon Warrior RTA Chart.xlsx

welcotar

Edit history:

welcotar: 2015-12-19 11:43:41 am

2015-12-19 02:28:13 am

I finally finished some hand axe routing: https://goo.gl/a5z2fi
video:

Different NPC movements map to slightly different paths, but all paths get to the same 16/9 axe knight fight. The jogs get them to sync up to this one good fight. The fight is two frame rules away, so you just stand on the swamp tile and hesitate "twice" as long as usual. I personally find that part easy. The open+close menu buffer is hard for me but ymmv.

Remembering the jogs might be hard, too, I dunno; the paths are pretty simple, one could probably just pick 1-2 that are most consistent and remember those.

If it's doable consistently, this should save a good chunk of gold grind, a trip to Rimuldar, and also be a pretty fast Axe Knight fight. Probably have to rely more on HURT for a few levels.