Post by TiKevin83

Posted: 10/24/2019 7:26 PM

Quote

AdituV

Player (74)

Joined: 8/26/2015
Posts: 70

My script is only for US right now, so yes. It would be pretty easy to adapt it for JP though, just needs a few memory addresses (gTextPrinters, gMain, maybe something else I forgot)

Posted: 10/24/2019 7:34 PM

Quote

merrp

She/Her

Player (18)

Joined: 7/31/2019
Posts: 56

In that case, you can find the address of gTextFlags (or any other global variable; static variables have to be temporarily made global, and the project re-compiled to show up here) in pokeemerald.map If frequent recompiling is an issue, you could make a script to parse pokeemerald.elf, I think all variables including static ones show up there. 0x03003014: gTextFlags for US

Posted: 11/19/2019 11:48 PM
(Edited: 11/24/2019 1:17 PM)

Quote

merrp

She/Her

Player (18)

Joined: 7/31/2019
Posts: 56

This has taken a while because of work and school, but I've started TASing the new route: https://www.youtube.com/watch?v=wzdV8BhJUQA Link to video After Norman, it comes out about a minute faster than my previous video, and that includes fighting an additional Solrock (for a total of 3), so I won't have to fight a Poochyena this run. It incorporates these changes: - Skipping Strength, thanks to research done by FractalFusion. This changes the route to skip going through Route 116 again entirely, instead backtracking through route 113/111. - Minor music fadeout abuse. The only location this really saved time for me was entering Meteor Falls. - Skipping naming Abra since I will need to name a Wingull, or something, instead later, as was done in the previous video. - Fixed a few small pathing errors. Since I started this branch from the Abra fight, there are 1 or 2 unchanged paths from very early on, that I didn't think were really worth changing. They're talking to Mom from the left, and talking to Norman from the right the first time. Since they're both pre-Ralts fight though, the timing is sort of synced to when the right Ralts can be encountered, so changing these paths will probably only save 4-8 frames at most, because I'll just have to run for longer in the grass to time the Ralts encounter *slightly* faster. TODO: - TAS to the end, to see how much faster the new route ends up being overall. - TAS a second branch, where I go to Verdanturf after Wattson instead of after Flannery, to see which is faster. - Other optimizations?? Lots of thanks to FractalFusion for research on a Strength-less route and calculating the necessary move order for fights, to have enough PP to beat Norman.

Posted: 11/24/2019 1:09 PM

Quote

merrp

She/Her

Player (18)

Joined: 7/31/2019
Posts: 56

Alright! Super happy with this, saved almost a minute and a half off of the previous full TAS. 206563 frames @ 59.73 fps = 57:38.42 https://www.youtube.com/watch?v=9WNBmiVMdo8 Link to video Changes: - Better understanding of the timing necessary to align memory for Glitzer Popping. Using the Route 102 spinner allows me to advance the RNG to get the frame I want faster. The memory alignment depends only on the encounter frame *and* the tile it occurs on--tiles close to the spinner add, I believe, an extra RNG cycle, while tiles farther away do not. - Improved some very subtle mistakes in the PC button press timing. - Tried a few different Box naming orders--it's counterintuitive, but deferring naming Box 2 until after picking up the Egg doesn't save time. - New payload that wasn't shown in the last video. More fun--tricks the game into thinking A is pressed so that the final real press happens much earlier. - As mentioned earlier, Strength is not picked up in this run. Due to this, Marshtomp's EV distribution was altered, and I didn't have to use a Qualot Berry to get glitch moves that behaved nicely! TODO: - TAS the branch going from Wattson->Verdanturf->Fallarbor, and see if it's faster - Possible improvements on both the bootstrap payload, and the code that is input. Making it slimmer by 1 instruction would save 2 frames. - Encounter "pathing"--basically finding the fastest way to an encounter--could use work. - Suggestions?? Newly changed, not in this video: - Ralts fight optimizations--talking to Norman from the left saves 32 frames If I can save another minute, it'll be 20 minutes faster than the current fastest completion!

Posted: 8/25/2020 9:29 AM

Quote

GoddessMaria

She/Her

Reviewer, Experienced player (849)

Joined: 5/29/2009
Posts: 514
Location: Hell...

...just going to leave this here.. Link to video

Current projects: failing at life

Posted: 2/21/2021 11:42 PM

Quote

Zowayix

Joined: 12/29/2007
Posts: 489

Out of curiosity, do any routing differences happen if the branch is timed from the very beginning of the save file instead of after beating the game? (E.g. Are there any places where it would be faster to prepare for the Battle Frontier before beating the game, at the cost of beating the game slower?)

Posted: 4/10/2021 11:22 PM

Quote

DanZC

Joined: 4/5/2021
Posts: 1
Location: United States

Hi. I'm trying to route a low pause TAS of Pokemon Emerald, and I'm having trouble wrapping my head around Glitzer Popping. Can someone help me?

Posted: 12/2/2023 4:08 AM
(Edited: 12/2/2023 4:55 AM)

Quote

jlun2

Skilled player (1705)

Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"

I'm not sure if this is the best place to ask, but is there a script that displays the current, and next several hundred frames of encounters? It has nothing to do with TASing; I just want to shiny hunt in R/S/E on emulator using either battle tower roaming bug, or normal wild battles. How are battle tower shinies determined? All 3 pokemon at the same time as the battle starts? What encounter method is it? Is it possible to have a trainer in battle tower with 2 or more shinies at the same time? To make this post less useless, I did search beforehand, and here's a bunch of past posts about gen 3 scripts, all in 1 place: Ruby/Sapphire thread: Pokemon IV/Nature/Hidden Power display by FractalFusion: Post #262235 Pokemon IV/Nature/Hidden Power display v2 by FractalFusionPost #351862 Repost of display v2: Post #431868 Pokemon Emerald thead: Pokemon data viewer by Metarkrai: Post #384715 RNG viewer for battles, encounters, catch rate, and even Pokérus by MKDasher: Post #457922 (dead dropbox link) gen3tools by merrp: Post #488195 Pokemon IV/Nature/Hidden Power display v3 by FractalFusion: Post #488732 Pokemon dialogue script: Post #488857 Pokemon Leafgreen/Firered thread: Pokemon IV/Nature/Hidden Power display by FractalFusion: Post #303383 Modified FractalFusion script by mkdasher: Post #353499 (dead mediafire link) Also, what is the current best route for 4 stars/corrupt them all (without ACE) in emerald? Is it still these? Post #431278 Post #405196 How feasible would it be to make a low optimized/almost casual but occasional savestate use run as a proof of concept given almost 0 knowledge of the memory of this game? Finally, is the new mail bug for R/S possible to be used in Emerald for faster than Pomeg ACE?

LLCoolDave wrote:

This seems to be competing for the "Worst looking TAS of 2011" Award.

Posted: 12/2/2023 4:37 AM

Quote

CasualPokePlayer

Emulator Coder, Judge, Experienced player (595)

Joined: 2/26/2020
Posts: 697
Location: California

jlun2 wrote:

Finally, is the new mail bug for R/S possible to be used in Emerald for faster than Pomeg ACE?

The Recycle + Knock Off glitch is usable in Emerald and hits party data somewhat similarly to FRLG, so it can be used to get a glitch mon and ACE. Of course, in practice this is TAS only as it requires Metronome since in Emerald you can't normally get any Pokemon with Recycle without trading or abusing Sketch.

jlun2 wrote:

Also, what is the current best route for 4 stars/corrupt them all (without ACE) in emerald? Is it still these? Post #431278 Post #405196

Given the above it's probably outdated by now, as Recycle + Knock Off would be a faster entrypoint to glitching (and at the very least could gateway into pomeg glitching).

Posted: 12/2/2023 4:56 AM
(Edited: 12/2/2023 4:57 AM)

Quote

jlun2

Skilled player (1705)

Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"

CasualPokePlayer wrote:

Given the above it's probably outdated by now, as Recycle + Knock Off would be a faster entrypoint to glitching (and at the very least could gateway into pomeg glitching).

Thanks. How would you get Recycle from Sketch in Emerald though without trading?

LLCoolDave wrote:

This seems to be competing for the "Worst looking TAS of 2011" Award.

Posted: 12/2/2023 5:09 AM

Quote

CasualPokePlayer

Emulator Coder, Judge, Experienced player (595)

Joined: 2/26/2020
Posts: 697
Location: California

jlun2 wrote:

CasualPokePlayer wrote:
Given the above it's probably outdated by now, as Recycle + Knock Off would be a faster entrypoint to glitching (and at the very least could gateway into pomeg glitching).
Thanks. How would you get Recycle from Sketch in Emerald though without trading?

Metronome

Posted: 12/2/2023 6:48 AM

Quote

merrp

She/Her

Player (18)

Joined: 7/31/2019
Posts: 56

jlun2 wrote:

I'm not sure if this is the best place to ask, but is there a script that displays the current, and next several hundred frames of encounters? It has nothing to do with TASing; I just want to shiny hunt in R/S/E on emulator using either battle tower roaming bug, or normal wild battles. How are battle tower shinies determined? All 3 pokemon at the same time as the battle starts? What encounter method is it? Is it possible to have a trainer in battle tower with 2 or more shinies at the same time?

Gen3Tools can display that info, if you input the RNG state into it; usually I have that and the cycle count on screen and I input one or the other. For battle tower shinies, this function is how they're determined: https://github.com/pret/pokeemerald/blob/master/src/battle_tower.c#L1633 They're generated one after the other, but depending on how much other code runs that frame, it's probably any method between 1, 2, and 4, depending on the number of nature rerolls.

Finally, is the new mail bug for R/S possible to be used in Emerald for faster than Pomeg ACE?

As CasualPokePlayer said it is usable in TAS; I am going to work on that after finishing FRLG. It'll involve maining Makuhita (For Knock Off) and maybe Zigzagoon (for Pickup Candies) with the goal of getting Makuhita to level 28 and getting to Fallarbor Town for the Metronome tutor (Abra can learn it so no problems there). Then I'll need to find a suitable target that I can corrupt to acquire a glitch species for ACE. The current TAS gets to Fallarbor at around 33 minutes, so adding in a few minutes for Makuhita and leveling, the final time could be sub-40, I think.

How feasible would it be to make a low optimized/almost casual but occasional savestate use run as a proof of concept given almost 0 knowledge of the memory of this game?

The decomps really help with this, basically every symbol has been mapped out. I'd say it'd be pretty reasonable to make a savestate run where you maybe do some RNG manip.

Posted: 12/2/2023 6:33 PM

Quote

jlun2

Skilled player (1705)

Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"

merrp wrote:

jlun2 wrote:
I'm not sure if this is the best place to ask, but is there a script that displays the current, and next several hundred frames of encounters? It has nothing to do with TASing; I just want to shiny hunt in R/S/E on emulator using either battle tower roaming bug, or normal wild battles. How are battle tower shinies determined? All 3 pokemon at the same time as the battle starts? What encounter method is it? Is it possible to have a trainer in battle tower with 2 or more shinies at the same time?
Gen3Tools can display that info, if you input the RNG state into it; usually I have that and the cycle count on screen and I input one or the other. For battle tower shinies, this function is how they're determined: https://github.com/pret/pokeemerald/blob/master/src/battle_tower.c#L1633 They're generated one after the other, but depending on how much other code runs that frame, it's probably any method between 1, 2, and 4, depending on the number of nature rerolls.
Finally, is the new mail bug for R/S possible to be used in Emerald for faster than Pomeg ACE?
As CasualPokePlayer said it is usable in TAS; I am going to work on that after finishing FRLG. It'll involve maining Makuhita (For Knock Off) and maybe Zigzagoon (for Pickup Candies) with the goal of getting Makuhita to level 28 and getting to Fallarbor Town for the Metronome tutor (Abra can learn it so no problems there). Then I'll need to find a suitable target that I can corrupt to acquire a glitch species for ACE. The current TAS gets to Fallarbor at around 33 minutes, so adding in a few minutes for Makuhita and leveling, the final time could be sub-40, I think.
How feasible would it be to make a low optimized/almost casual but occasional savestate use run as a proof of concept given almost 0 knowledge of the memory of this game?
The decomps really help with this, basically every symbol has been mapped out. I'd say it'd be pretty reasonable to make a savestate run where you maybe do some RNG manip.

Thanks! I was looking at https://github.com/pret/pokeruby/tree/master/data/scripts Since the battle tower bug only applied to R/S, and I thought Enerald's frontier tower would be too wildly different to check. If the mail bug works similar to Pomeg for corruption, does that mean Fr/LG 4 stars is feasible too? :o

LLCoolDave wrote:

This seems to be competing for the "Worst looking TAS of 2011" Award.

Posted: 12/15/2023 3:17 AM

Quote

FractalFusion

Editor, Skilled player (1938)

Joined: 6/15/2005
Posts: 3246

merrp wrote:

It'll involve maining Makuhita (For Knock Off) and maybe Zigzagoon (for Pickup Candies) with the goal of getting Makuhita to level 28 and getting to Fallarbor Town for the Metronome tutor

In Emerald, Pickup is level-based. Zigzagoon / Linoone is not available at the necessary level to get Rare Candy until after Norman, unless you level Zigzagoon to Lv 21.