RNG Manipulation in FireRed/LeafGreen: Wild Pokémon Supported in RNG Reporter 9.93

[Arseus]

GoldenBanana, sorry for the delayed reply, though as you correctly guessed, everyone here is very busy! The reason that we are unsure as to when exactly the PID is set for Raikou/Suicune/Entei is the IV glitch which you have recognized in your post; there is no need for the collective community to know when those Pokémon specifically are generated. For the sake of completeness however, I will edit your research into the FR/LG Legendaries guide when you find the line of text when they are generated.

Now, I recognize that you have already addressed the IV glitch in your post, and are well aware that the posters here may try to dissuade you from continuing further. However, I still feel obligated to try to urge you away from this specific test. You are likely well aware that you can get a shiny Entei from HeartGold/SoulSilver, especially since you can get your SID from the Red Gyarados, so I will assume that you have already decided against doing so, and I will not pursue this further. Nonetheless, as you may notice from the various posts in the thread, it is still difficult to hit your seed and frame in FireRed/LeafGreen, and even then, the end result will be an Entei with a low Attack IV and a load of 0s as its IVs. In short, I am convinced that the amount of effort you would put into this would outweigh the eventual payoff.

In fact, I will take this opportunity to honestly say the following about FireRed/LeafGreen RNG manipulation: with the advent of Generation V Dream World abilities and a "nerfed" SelfDestruct for Mewtwo (in addition to the new "physical" Psychic move which runs off its Special Attack), it has never looked less appealing, especially given its purported difficulty. I say this not to discourage any of our current researchers, but merely to provide a realistic assessment of the situation. I would still like to see this through to the point where our users can RNG abuse for Pokémon in FireRed/LeafGreen with requisite proactivity, and as such, I commend and encourage the efforts of all our researchers.

 

[Nix_Hex]

Thank you GoldenBanana and Arseus for bringing this up again. My interest in this project, while constantly on my mind, has dwindled, for several reasons. My heart sunk when I read about the nerfed mechanics of SD/Explosion - any successful attempts for a near flawless Mewtwo will only be useful in Gen IV wifi battles, which will all but disappear. A 200 base power non-stab self-ko off of an Attack stat not much higher than that of Aerodactyl is pretty worthless. Also, the information for seed generation is incorrect, sending the very idea of tempo/beats the way of the dodo (sorry for the hype). I have made a few videos showing the "real" creation point, but I can count the number of users who would be interested on one hand. Unfortunately, rng abuse in FRLG will be just as proactive as emerald: find what you can with what you're given, except with these games you can get a small range of common seeds, giving you many more options. I likely will not work on this in the near future, but if I find anything out, I will let it be known.

 

[GoldenBanana]

I understand that it seems like such a waste of time and effort to continue this research since anything that can be obtained from it is easily outclassed because of the later generations. It is especially frustrating that 5th gen has nullified a previous incentive, SD Mewtwo, with the new mechanics.

I know that it's pointless to abuse for the dogs in FRLG, since it is easier and more profitable to do so in HGSS, but I'm only working on Entei because I had already started the game with bulbasaur before I learned about the research; if I had started with squirtle, I would be working on Raikou, despite the fact that I already have a shiny, Timid, 31/31 sp.atk/spd IV, HP Ice 68(or something) Raikou running around in my HeartGold. I'm only interested in this research because 3rd gen is my favorite of all the Pokemon generations, so I want to make the most out of it.

NixHex, when you say that the information for seed generation is incorrect, do you mean the part where the game starts seeding after the intro up to four frames after start is pressed, and/or the part where 65535 seeds cycle every 3ms at the start screen? I know that you've been able to hit 4 or so different ranges of seeds commonly, but wasn't that consistent with the seeding information we had, or am I missing something? Also, I'm interested in seeing these videos showing the "real" creation point of the seeds. Where can I see them, or can you give me the gist of the information?

Now, I'm aware that what I'm researching won't interest very many people, but to avoid spamming this thread with findings of my research, I'm going to do what NixHex did with his post and turn my previous post into a log of my findings and methods while trying to RNG abuse for Entei. But before that, I'd like to bring up an issue that I'm experiencing with my research. Since Entei can only be identified by his HP IV, and sometimes his Attack IV, I have to make sure that I'm certain about what seed I'm on, so I use two pokemon to check my seed. However, when I use the FRLG Seed Finder application and plug in the seeds I get from the "IVs to PID" tab on RNG Reporter, I either get no seeds at all, or I get seeds where only the pokemon's IV spread whose seed yielded that 2-byte seed is found when I plug the 2-byte seed into RNG Reporter, and that of the other pokemon is not. Does this mean that I somehow switch to another seed without restarting the game, or maybe I don't have a very good grasp of how seeds work in this game? Either way, I can't continue until I figure this out. It's frustrating enough that it takes at least 45 minutes to do the entire process once.

 

[Arseus]

NixHex, I too would be interested in knowing the true creation point of the seeds, if only so that I may amend the first post with the accurate information. I understand that you are rather discouraged at this point both after discovering the fifth generation and the true creation point of the seed, which makes a large portion of your recent research—by your own admission—largely obsolete. If you do not wish to continue with this research, I understand fully, though I would like to also take this post to sincerely thank you for the research you have done, in addition to the intent behind the research that is no longer in the first post. Some of your work is still very usable, so it is not as though all was for naught.

GoldenBanana, I had a feeling you had some sort of affinity for the third generation, which is why you would put up with the legendary beasts (FR/LG, Colosseum) which call it home. Just a note on your methods, Entei's Attack IV will always be less than or equal to seven (or some such number as I recall), which means that you should not really be using it to identify your roamer. Now, concerning your problem with seeds, I cannot admit to actually using NeoSyrex's program in a test setting myself, but on the surface (perhaps someone else like NixHex can correct me), I cannot detect your error. I assume there are a number of different seeds that come up in the IVs to PID function, are there not? Have you tried each one, and the different methods that are provided as well? As you may be aware, the PID generation of wild Pokémon is rather fickle in what method it chooses.

As for the presentation of your research, I would encourage you to make a new post if you stumble upon something that may be relevant to the masses; if you would be making a double post, edit your most recent one, and shoot me a PM so I can take a look at it and comment if I can. I don't think I will necessarily be including your "Entei info dump" in the original topic if only because it is, as you have identified yourself, very niche. Despite my presence at the moment, I am actually very busy (I am right now making some use of downtime), and as such cannot necessarily respond to this thread as quickly as I would like.

 

[Nix_Hex]

Hehe fine, I'll speak! I am on a cell phone and will be for the next 6.5 hours so excuse grammar or punctiuation errors. Here was my process:
I started a new file in a rom of LeafGreen, but I was unable to save it due to some battery error (it is in fact a rom). So, I started the rom anew and watched the memory address supplied by flovv. The address chose its seed a few "moments" after Venusaur's cry. I highly doubt this time changes when there is a save file but I am unable to find out, due to the reason stated above. I believe the frame (not the seed) I hit was also the trainer ID and probably determined the iv's of my starter (Charmander), so this might mean that only certain TID's are compatible with the starter's spread. That's a whole other road that I'm not exactly interested in but Toothache might like it. I will narrate the videos tonight and put them on my Youtube channel.

GoldenBanana, the process for finding your seed is as follows: capture a Pokemon at a high enough level to get the exact IV spread. Next, plug the spread and nature into the IV to PID applet and choose a Method 1 seed (there might be two). Plug it into NeoSyrex' program and if a 2 byte seed comes up, then you were successful.

I'll be updating this post since it is on the most recent page. As a side note, I start grad school in March and will be extremely busy. But, I will try to include this sort of research into my thesis (I'll find a way) to give me an excuse to take breaks from homework to play Pokemon.

Arseus, I appreciate your gratitude through these past few months. I hope this new research leads to more success.

 

[Zari]

Whether or not the mechanics change will change how mewtwo is played, selfdestruct is still a potent move. It still has the use of letting a new threat come in unscathed, while dealing at least a little bit of damage at the same time.

The other thing to consider is SD usage in doubles; afaik (correct me if I'm wrong), the new psychic move is single target only, while SD keeps its "entire-field" attack range. Yes, the usage in singles might decrease, but imo it'll still see at least some usage in doubles (or triples come VGC 2011?).

---

Back on topic, even if the tempo/beats are out, we still have the other method I discovered with skipping the intro for consistency; THAT at least produces (somewhat) consistent results.

If you want me to Arseus, I can edit into this post a more formal, step by step guide for you to link to the OP. However you reply is fine with me.

EDIT: Okay here's a basic rundown of how the method works:

----

Important Info

As mentioned in one of NixHex's earlier posts, this is a reactive-proactive method; you find groupings of seeds that you can hit, then search for spreads within those ranges that are of interest. NixHex and I have hit similar groupings of seeds independently, leading me to the conclusion that it *should* be possible for any user to hit those seeds/the spreads within them as well.

The [combined] data we have collected is as follows:

Seed [Hex] groupings (an "X" indicates a variable value)

54XX (XX usually ranges from the values c0 to d4/d5)

30XX (XX: 82 to a0)

0bXX/0cXX (XX: ef to 07 [latter value has 0c at the forefront])

9eXX (XX: 0c to 17)

There are other groupings, but these are the most common. Now onto the method itself!

The "Skipping" Method

Basically, what we need to do is skip the intro and charizard/venusaur "flash" sequence entirely. We do this by button mashing, but it requires some timing as well.

I don't have the exact/close approx. time length nailed (working on this as I find the time), but you need to begin button mashing the "A" button after the credits/nintendo junk, but directly before the screen opens up from the center with a deep blue (this is the background screen which the Game Freak star passes over).

Directly after the introduction finishes (or would finish, if you let it run its entire course) there is the zard/venu "flash" as I like to call it, where the pokemon's sprite appears white, then forms a grey version of itself, and then becomes full color after that subsequence finishes. What we want to do is press "A" fast enough so that we skip this "flash" entirely.

The final part of this method is to press "A" fast a third time, after skipping both of those sequences. If done correctly, the cry of venusaur/charizard should sound within 1 to 2 notes of the pokemon theme's beginning.

Warning: this may take a lot of tries to get correct; the desired time length between button presses is ideally zero, but, since that's physically impossible, you want to aim for as small as you can make it. The smaller the time length between button presses, the more consistent results you will get.

That's basically it for the method. After all of that is complete, do the usual by catching mewtwo (or another high level legend/mon), finding the PID via RNG reporter, then inputting that seed into NeoSyrex's FRLG Seed Finder.

Notable spreads within the given groupings:

For Mewtwo:

Seed: 54cc, F: 1838, Naive 27 / 28 / 30 / 30 / 27 / 31

Seed: 3081, F: 22538, Hasty 30 / 30 / 26 / 30 / 30 / 31 -- note, I have not been able to hit this seed yet, though I have hit ones around it many times.

Seed: 9e0c, F: 25054, Naive 28 / 31 / 31 / 27 / 23 / 31

Seed: 0bf0, F: 30980, Hasty 31 / 30 / 25 / 31 / 25 / 31

Seed: 0bff, F: 20091, Hasty 28 / 31 / 30 / 27 / 30 / 31

Seed: 0c00, F: 6208, Naive 29 / 31 / 24 / 29 / 31 / 31

For those who might like to play ADV sometime again:

Seed: 54c6, F: 8631, Calm 31 / 31 / 30 / 31 / 28 / 31

Seed: 3082, F: 6403, Naughty 31 / 30 / 30 / 30 / 22 / 31

Seed: 9e14, F: 35262, Naughty 30 / 29 / 27 / 30 / 27 / 31

Seed: 0c02, F: 11279, Rash 30 / 30 / 31 / 27 / 31 / 31

Seed: 9e10, F: 63361, Jolly 28 / 31 / 29 / 1 / 31 / 31

---


There's most likely more, but that's just an example of what you can find with a few groupings of seeds. All but the last jolly one are under 10 minutes; most are under 4. Anyways, I'll leave the rest of you to keep digging. Just be sure to post if you find stuff/have successes with this method!

 

[Arseus]

I look forward to any further research you have NixHex, and I will be certain to link to your YouTube videos in the main post when they are uploaded. Your new findings are quite intriguing, and also reveal the unfortunate reality of how difficult it is to get information on different game mechanics.

Zari, I would be happy to include a link to or repost of your mini-guide in the original post. As of this edit I have now done so; your post has been quoted, and said quote also contains a link to the original post, as per most links I have in the topic.

 

[Nix_Hex]

Here is the video, as promised.
http://www.youtube.com/watch?v=9gBkheUj8KE
The volume is a little imbalanced, especially when the game comes in. So, basically, you're seeing the seed copied from the timer to its final address at 0x0200020.

 

[GoldenBanana]

It looks like the seed is chosen as soon as the start screen fades out, or as soon as the continue screen begins to fade in. If someone can figure out how many frames after pressing 'A' or 'start' at the start screen the seed is chosen, then the beat/tempo method still has chance, right? It looked like around 1.5 seconds, so something like 90 frames? Anyways, thank you for sharing the video NixHex.

In other news, I edited my post on the previous page yesterday and put up my method of finding out when Entei is generated, if anyone is interested. Perhaps someone can look it over and critique it?

 

[Nix_Hex]

GoldenBanana, if you use www.metalkid.info, you'll have a much easier time finding your captured Pokemon's IVs - I usually use between 2 and 4 Rare Candies for Mewtwo at level 70, so I don't think you'll be using more than 6 for a level 64 Pokemon. This goes for Entei, who I believe is level 40 (?), so I imagine you'll be using around 9 Rare Candies (plus you only have to narrow down HP and Attack). Also, you seem to be having a hard time with the seed finder. I'll make another video (without narration) showing this whole process. I don't know why it hasn't worked for you though.

Unfortunately, I don't know if the beat tempo method is reliable any more than is spamming A at the Gamefreak star. Believe me, I thought about finding out that interval, and it may be as easy as starting a stop watch when you press A and stopping the stop watch when the seed is chosen, calibrating it until we get it just about right.

 

[Arseus]

Okay, NixHex, I have replaced the old part of mattj's post with the link to the relevant YouTube video. I will continue to monitor this topic to the best of my ability, and will update with any other videos you may post.

 

[GoldenBanana]

Thanks for looking at my post NixHex. Metalkid's calculator makes the process of finding everyone's IVs a lot faster. Looks like I wasted my time collecting so many Rare Candies. =/ I'll make adjustments to my Method the next time I record any Trial Data.

I've been able to deduce my initial seed a couple of times, but sometimes the Seed Finder gives me several matching initial seeds, so it is still somewhat difficult to accurately find my initial seed. I've already posted one of my trials on the previous page, but before I post anymore, I want to clear a few things up:

(1) Is the Cerulean Cave a quiet or noisy area? I need to know this because the only way to eliminate some possible initial seeds is to find out the approximate frame at which I encountered the pokemon. I have a suspicion that it is, but I need a confirmation.

(2) When Entei's Attack IV is less than 7, was it generated normally, meaning that the Attack IV in RNG Reporter on that frame will be whatever Attack IV Entei has? Or should I disregard his Attack IV altogether?

 

[Nix_Hex]

(1) Is the Cerulean Cave a quiet or noisy area?

Zari and I have done extensive research on Mewtwo's basement, and it is for sure quiet. As for the rest of Cerulean Cave, I haven't personally tested this, but it's probably uniform.

(2) When Entei's Attack IV is less than 7, was it generated normally, meaning that the Attack IV in frame will be whatever Attack IV Entei has? Or should I disregard his Attack IV altogether?

This probably isn't hard to test. Keep track of approximately the time it takes to generate Entei so you can use its seed to find the spread (using only the hp iv and nature). Something else to consider is if the RNG is *supposed* to generate a Attack IV greater than 7 on a certain frame, does it automatically roll it down to 7 or possibly a number below seven? It's probably the former, but I don't think the latter would be too much of a monkey wrench.

 

[Toothache]

Can I suggest for future videos you use RAM Watch, it is much less confusing to the average person. Also, it would be helpful to note when the ID/SID is generated (at the name selection screen) since that is important too.

Oh, and in order to remove the 1M circuit board message, set the following options in VBA22:

Options -> GBA Emulation -> Real Time Clock active
Options -> GBA Emulation -> Save Type -> Flash 128k

 

[Nix_Hex]

Can I suggest for future videos you use RAM Watch, it is much less confusing to the average person. Also, it would be helpful to note when the ID/SID is generated (at the name selection screen) since that is important too.

Oh, and in order to remove the 1M circuit board message, set the following options in VBA22:

Options -> GBA Emulation -> Real Time Clock active
Options -> GBA Emulation -> Save Type -> Flash 128k

Thanks, Toothache. Yeah, I was using the "normal" VBA when I made that video, before you told me to get the enhanced versions. The reason I didn't try to show ID generation is precisely because of that 1M circuit board error. I have downloaded the lastest VBA-rr and have been toying with Emerald (which, interestingly enough, uses the same internal clock to generate the ID/SID that FR/LG use to generate the seed, and, I'm assuming, the ID/SID). Hopefully, I'll have some time to get back into Emerald so I can mess around with the ID some more. Hoping to get 00311, hehe.

 

[Nix_Hex]

Sorry for interrupting and changing the subject, but in this post, I'll put the stuff I find about ID and SID creation in FrLg.

-Seed 1 is created after pressing A on the Venusaur/Charizard Screen
-Seed 2 is created when you press OK after you have typed your name.
-By saying No, and therefore retyping your name, you can redo the seed.
-Seed 2 is your TID(converted to decimal)
-Seed 2 appears to overwrite Seed 1(or that's my explanation for it)


If you guys already knew this, tell me now so I don't make a fool of myself. Also, a quick double check, if I'm on VBA, I can check the seed using Memory Viewer. If I recall correctly, it's location is 02020000 right?

I'm sorry to double post and answer this so much later, but you're right, the address for the seed is 02020000 copied from the clock address in my video. I've been messing around in Emerald (the visible ID is created using the same clock address as FR/LG, right after the character's gender is chosen, but I cannot for the life of me find where the Secret ID is generated! If it is copied from a clock, that would be very interesting to know! I'm using a Gameshark code that reveals the Secret ID in the Pokémon's Info, and I noticed that the Secret ID exists at 02020000 and several other addresses some time before you receive the starter (or as in the case of Emerald, battle with that starter). It is definitely independent of the Trainer ID, since I have used a save state and received different SID's with the same TID.

 

[mattj]

GoldenBannanna when Steve's Seed Finder gives you more than one result just look at the frame. 99 percent of the time, as long as you're encountering your target within a minutes consistency you will have frames that are pretty close to each other. The incorrect seeds will have unreasonable frames.

 

[Zari]

Hey just posting to say I did some more testing on mewtwo, and I've hit an extension on one of the groupings; the 54XX grouping can go up to at least e0 for the XX, which means there is a fantastic timid spread now available:

Seed: 54de, 42197F (11:43.28 :/ ), Timid 28 / 31 / 30 / 31 / 30 / 31

Just thought it might be interesting to note that one.

I'll edit this post if anything else interesting comes up.

 

[GoldenBanana]

GoldenBannanna when Steve's Seed Finder gives you more than one result just look at the frame. 99 percent of the time, as long as you're encountering your target within a minutes consistency you will have frames that are pretty close to each other. The incorrect seeds will have unreasonable frames.

Thank you for your input, but here's the scenario I'm working with:

I start the game at One Island Pokemon Center, and have to travel all the way over to the basement of Cerulean Cave, catching Entei along the way. This ends up taking 5 or so minutes at the very least. So when I check my seed, I'm not comparing a small frame to a very large frame, eliminating the large one; I have to compare two large frames (10 min+)that are only 1-3 minutes apart from each other. Also when I compare the frames to the external timer I use, they don't match. Specifically, the frames are usually higher.

I've only tried this experiment using a timer 3 times, so I'll have to do a few more runs to see how far off the frames are. Obviously this has to do with noisy/quiet areas of the game. I've also been away from this project for a while, but I'll see if I can do a few more runs today.

I appreciate the input though.

 

[mattj]

Oh. Just set up another timer (anything would work) and DON'T catch Entei till the timer runs out. Like. Have a mon 1 level below him and some repels and don't start running around till the timer runs out. That way your frames will be decently close to each other and you'll know what's the right seed, and what's not. You could probably catch him within 30 seconds of each other then, and the frames would be reasonably close.

[edit]
or wait, do you know what area he will be in? i don't think I've ever went for Entei in FR/LG.

 

[GoldenBanana]

Oh. Just set up another timer (anything would work) and DON'T catch Entei till the timer runs out. Like. Have a mon 1 level below him and some repels and don't start running around till the timer runs out. That way your frames will be decently close to each other and you'll know what's the right seed, and what's not. You could probably catch him within 30 seconds of each other then, and the frames would be reasonably close.

[edit]
or wait, do you know what area he will be in? i don't think I've ever went for Entei in FR/LG.

Oh, that's not a problem. Entei is generated after you fix Celio's machine (not when you encounter him, thank God). The whole point of this research is to see exactly when during the sequence he is generated. I've listed my method in my last post on the previous page.

 

[Zari]

Well people, I've got good news; I've discovered more groupings of seeds! But before I start on that, I'll give you the data first (still working on it, so this post will be edited with more data in the future)

By groupings:

9eXX [02 - 2d]

9e02 * 2
9e03
9e0a * 2
9e0b * 4
9e0c
9e0d
9e0f * 4
9e13 * 8
9e14 * 10
9e18
9e1c
9e1d * 2
9e29 * 2
9e2d

54XX [c2 - f3]

54c2 * 2
54c3
54ca * 2
54cb * 6
54cf * 2
54d2 * 2
54d3 * 2
54d4 * 10
54d7
54d8
54db
54de * 2
54e0
54ed * 2
54f3

30XXv1 [83 - a8]

3083
308a
308b * 5
3093
3098
309c
30a0
30a8

30XXv2 [3d - 5a]

303d
3045 * 2
3046 * 2
3047
3048
304b
304c * 5
304d * 2
304f * 2
3050
3056 * 2
305a

e7XX [42 - 72]
e742
e749
e74b * 8
e74f * 5
e758 * 3
e75c * 2
e75d * 3
e769
e771
e772 * 3

I'll be adding more in later as I find the time.
---

One other thing that I managed to do was hit the same seed AND frame twice with only emloop. I think that's at least some indication that RNG manipulation in FRLG is entirely possible, as difficult as it might seem.

Now back to more resetting :)

 

[Nix_Hex]

Well, it looks like the 9exx grouping is definitely the most dense, which I've witnessed on my own accounts. Some interesting things have happened to me while resetting for Squirtle on an emulator were that I hit seeds within some of those ranges (I used RAM watch) without button mashing, just waiting arbitrary amounts of time. This is totally baseless but I think that the range of seeds may be determined by one clock (perhaps at 60 Hz), and the seed you hit within that range is determined by the internal timer as discovered by flovv. I'll have some time this evening to research this more thouroughly, that is if I'm not engorged in Smogon's new PO (yuck) server. What I'll do is run the emulator at 1/60 and see if seeds from that range pop up while the clock is running.

 

[GoldenBanana]

Okay, so I've added a few more trials to the data in my post on the previous page, and it looks like Entei might be generated when you press 'A' at the last line of text in the conversation with Celio (which was a bit obvious, but still, it's best not to assume). However, this is not guaranteed yet, at least not until I've gotten shiny Entei to appear.

That being said, I will start trying to hit a desired seed soon. Since Zari has had some success with emloop, I'll try that method first. If that fails, I may have to try a reactive method to obtain a useable seed, which shouldn't be too hard since IV's aren't really a big concern, but I'd like to avoid that method nonetheless.

EDIT: I was just trying out the Skipping Method and I hit 54cc on my third try! Unfortunately, I was aiming for a much higher frame, so I missed frame 1838. Anyways, currently I'm calibrating the time it takes from when I start pressing 'A' to use the skipping method up until the seed is generated so that I know how much time to add to emloop's second timer.

EDIT2: Alright, I think I've made all the preparation I need, so I'm going to start abusing for Entei tomorrow. It doesn't look like I'll be able to check my seed before I generate Entei since changing routes/entering the center switches the frame advancements from quiet to noisy, and I don't know if I'll have much luck hitting my frame that way.

So instead, I'm going to try to hit my seed using a combination of emloop and a metronome, rather than using the skipping method, because the seed I'm targeting has a low enough frame to tolerate, and it's not within the groupings of seeds found through the skipping method. Zari, I know you said that you were able to hit the same seed using only emloop. Were you using the skipping method as well? If not, please tell me how you did it, because any sort advice would be really helpful when I take on this impossible task.

 

[GoldenBanana]

Spoiler: No reason for real excitement

I know this is a double post, but it's been quite a while since the last one. As I've stated before, I'm not using the skipping method to hit my seed. Instead, I used the beat/tempo method to determine the length of time needed to wait at the 'PRESS START' screen, and entered that time into emloop's first timer.

I also attempted to calibrate the time it takes between pressing 'A' at the 'PRESS START' screen until the seed is chosen. I did this by entering an arbitrary amount of time, 10 seconds, into emloop's first timer, and my target frame in the second timer. I would then reset/start the game and the timer, not necessarily at the same time, skip to the 'PRESS START' screen, and press 'A' as soon as the first timer counted down to 1 milisecond. I then collected Eevee at Celadon City once the second timer hit zero. After multiple attempts, I found that I was consistently 118 frames off.

With this information, I subtracted 118/60 from the first timer (the one found through beats and a tempo, not 10) and added it to the second timer. Since I can only find how far off I am from my target frame, I am assuming that the frames start rolling as soon as the seed is chosen. With these adjustments, I went to One Island to attempt manipulating Entei. After generating Entei, I checked my seed by catching any wild pokemon on Treasure Beach and leveling it up to 50. After finding my seed, I went to the route above Vermillion to search for Entei. With only Entei's Nature and HP IV, I was able to accurately check the frame he was generated at every single time. Therefore, Entei is definitely generated after the last line of text in the conversation with Celio.

Once I hit my target frame three times in a row, I stopped looking for Entei, and started to reset after I found that I didn't hit my seed. I have not hit it yet, but I have come close twice. In my last attempt, I hit cea6, and my target seed is ce12. That's 148 seeds off. The closest I have come is 127 seeds off, at ce91. To find the time required to wait at the 'PRESS START' screen, I used 22 beats at 63 BPM. This helps me hit around cf3d, which is 299 seeds off from my target.