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

[Arseus]

Hello again chrish! I am pleased to see that you have posted again, and with some new findings at that! I am sorry that I did not respond earlier, but I now have the time to give you some feedback.

Your Mewtwo would be Method 1, as Mewtwo is one of the few Pokémon that is consistently generated via said method, as outlined in the original post. Now, this is excellent that you were able to produce some results... this was indeed a cartridge run, correct? If so, I am pleased to see that we have now had a cartridge player hit the same frame twice; clearly we are making some progress. Now, I have a few constructive comments concerning this latest experiment.

Originally Posted by Fat chrish

Beat Count (cumulative) at each step (game event)
------------------------------------------------
0 Switched the game on in sync with the metronome
70 Pressed the Start button on the title screen
80 Pressed A on the Continue screen
160 Pressed A in front of Mewtwo
170 Pressed A after "Mew" is displayed
190 Pressed A after "Wild MEWTWO appeared!" is displayed

(...)

3) We will vary each interval in turn in order to try and determine which intervals are relevant.

Please allow me to save you a bit of time by saying that the time at which you press A when "Wild MEWTWO appeared!" is not relevant. The game chooses the frame (and by extension, the PID of your Mewtwo) as soon as the screen fades to black before the battle. As such, the last step is irrelevant to RNG manipulation, so you do not have to worry about it hereafter. On a similar note, I think it may be significant to note that the first three times would presumably have an effect on the seed (0 to 80 beats), whereas 160 and 170 would concern the advancing of the frame.

If I may, I would also like to make a suggestion as to a research question to sort out. If you can, try the following: go into your game (no special steps required) and wait for some amount of time (this may be completely arbitrary). Then, save your game and shut it off. Then, proceed with your experiment once again, and see if you are able to hit the same seeds as you were previously. If the seeds are the same, we may be able to eliminate the playtime of the file as a variable ("time in-game" may refer to the way the frames advance; loadingNOW's old comment on the savegame affecting the seed could be outdated). If they are different after a substantial amount of trials, then we have good evidence that the playtime is indeed a factor that influences the seed.

As something similar to try, you may wish to toy with the button presses on the title screen/menu to see if we can confirm NeoSyrex's theory of button presses affecting the seed. However, these other things may be tested after you are comfortable with the "vanilla" test results, and whenever you have the time.

I would like to thank you again for your enthusiasm concerning this project, and for the amount of time that you have likely been putting into this research. Thank you for your research, and I eagerly await your continued efforts!

[Newts]

I'd first like to congratulate you for achieving the same Seed/Frame, I too am playing with this, and i liked your metronome idea so much I'm also using myself with a beat count I feel comfortable with.

What I have found though is that the RNGReporter that you are using, is not really the best for GBA seeds. I'm capturing Zapdos as he was the easiest for me to get to on a speed run and after 20 captures I also ran the results through the RNGReporter PID finder. What I noticed was, every single spread was listed on frame 1, and this didn't seem correct at all. I then used PokeRNG with the same spreads and found I had a very different result.

I tested this with your duplicate Mewtwo and got this:
c7c68e22 (CAREFUL, ability 0) 29/0/26/16/4/12 (Seed: 1c0d06fa Frame: 1259) HP FIGHTING 32

PID is identical but with a completely different seed, and a believeable frame number. Hopefully someone who knows a litle more can fill me in as to why this is happening.

What I'm doing with Zapdos is getting a log of seeds hit using this metronome technique to see what kind of pattern appears, if any. Then I will let a full minute pass, save the game, and repeat the process to see what has changed by changing the gametime clock.

Once I am done with all this I'll post my results and link to them, hopefully the community can make use of them in ways I cannot.

[Arseus]

Newts_Ute, it is great that you are beginning to play with the RNG, and that you have been getting similar results to chrish. As for your concerns, I don't have a conclusive explanation for what is going on, but mingot has offered the following explanation to me based on the information in the thread:

Originally Posted by Fat mingot

RNG Reporter does not give you an "initial" seed in the IVs->PID function, nor does it try to or claim to. It gives you the RNG seed so that the subsequent 4 or 5 *depending on method* calls to the RNG produce the frame in question.

I have no idea about what pokerng gives because I did not write it. But I am pretty sure it has no facility to reliably give FR/LG initial seed.

I have contacted Wichu concerning PokeRNG's role in all this, but I suspect that only NeoSyrex's program would provide the most accurate look at the seed in FR/LG. NeoSyrex, if you are reading this, and able to post your basic DOS program, that could be very helpful at the moment. Hopefully I will hear from Wichu soon; I am sorry I cannot be more useful right now!

[Newts]

Thank you Arseus for the quick reply.

Originally Posted by Fat Arseus

I have contacted Wichu concerning PokeRNG's role in all this, but I suspect that only NeoSyrex's program would provide the most accurate look at the seed in FR/LG.

I've been checking frequently to see if this has been posted, my knowledge concerning these programs is somewhat limited so making my own is proving harder than expected. I guess I'll just keep collecting my PID information til I get my batch program running correctly or someone beats me to it.

Just for clarification, at the moment I'm using a spread sheet with the following formula, please correct me if I'm wrong:
PID - 6073
---------- = Seed
41C64E6D
Then having it repeat itself til I find a 2byte number as suggested by NeoSyrex.

[yayece]

Originally Posted by Fat Newts_Ute

Just for clarification, at the moment I'm using a spread sheet with the following formula, please correct me if I'm wrong:
PID - 6073
---------- = Seed
41C64E6D
Then having it repeat itself til I find a 2byte number as suggested by NeoSyrex.

No, the PID is made up of 2 separate RNG calls. This won't really do anything for you. You're best off using the "seed" that RNGreporter gives. Even then, I'm not entirely convinced that you can ever be sure of the initial seed, 2-byte number or not.

chrish: Do you have the seeds for all of your Mewtwo as reported by RNGreporter as well? If you can post those, we can examine them and see if any non-identical Mewtwo had the same initial seed as well.

[Arseus]

On the subject of RNG Reporter, PokeRNG, and seeds, I have received a PM from Wichu that has essentially confirmed my suspicions on the supposed third gen seeds from the test results:

Originally Posted by Fat Wichu

PokeRNG's seed is only for the 4th gen games, which seed based on the DS clock. FR/LG's seed, according to NeoSyrex's post, is always between 0 and 255 (0x00000000 to 0x000000FF). You don't want to use RNG Reporter's IVs to Seed function, as that only works for 4th gen games, where the rules for generating the seed are different.

Basically, the programs out at the moment (PokeRNG and RNG Reporter, mainly) aren't designed for FR/LG, so will never give you the correct seed :\

So until we get NeoSyrex to release his program, or someone else codes a problem that serves the same purpose, we won't be able to get the exact seeds for FireRed and LeafGreen with the programs we have right now.

[mingot**]

Originally Posted by Fat Arseus

So until we get NeoSyrex to release his program, or someone else codes a problem that serves the same purpose, we won't be able to get the exact seeds for FireRed and LeafGreen with the programs we have right now.

Well, here is the formula for the reverse RNG, which can be used to work backwards looking for an initial seed:

        public UInt32 GetPrevious32BitNumber()
        {
            seed = 0xeeb9eb65 * seed + 0xa3561a1;

            return seed;
        }

[megamonk4]

I would like to help out finding out how to abuse FR/LG.

One question that I have is if you could provide an example of how to reverse RNG the seed using the formula you've provided.

[yayece]

Originally Posted by Fat megamonk4

I would like to help out finding out how to abuse FR/LG.

One question that I have is if you could provide an example of how to reverse RNG the seed using the formula you've provided.

Well, take Emerald's RNG for example. You know that the initial seed is 0. Stepping through the PRNG, you get the following 32-bit values:

0x00000000
0x00006073
0xe97e7b6a
0x52713895
0x31b0dde4
0x8e425287
0xe2cca5ee
0xafc58ac9
0x67dbb608
0xfc3351db
0xef2cf4b2

Run it through mingot's equation, and you get the same sequence in reverse.
i.e., previous_value = (0xef2cf4b2 * 0xeeb9eb65 + 0xa3561a1)%0x100000000
That will return 0xfc3351db. Run 0xfc3351db through the equation, and you'll get the number before that: 0x67dbb608. Keep on going, and you're just walking through the PRNG loop backwards.

All that said, this is very nice to know. Still... I am not convinced of NeoSyrex's claim that the initial seed is always a 2-byte value. (0x0000 -> 0xffff) For example if you take chrish's Mewtwo example, the supposed "initial seed" is nearly 200k RNG clicks prior. That doesn't make any sense...

[megamonk4]

Originally Posted by Fat yayece

Well, take Emerald's RNG for example. You know that the initial seed is 0. Stepping through the PRNG, you get the following 32-bit values:

0x00000000
0x00006073
0xe97e7b6a
0x52713895
0x31b0dde4
0x8e425287
0xe2cca5ee
0xafc58ac9
0x67dbb608
0xfc3351db
0xef2cf4b2

Run it through mingot's equation, and you get the same sequence in reverse.
i.e., previous_value = (0xef2cf4b2 * 0xeeb9eb65 + 0xa3561a1)%0x100000000
That will return 0xfc3351db. Run 0xfc3351db through the equation, and you'll get the number before that: 0x67dbb608. Keep on going, and you're just walking through the PRNG loop backwards.

All that said, this is very nice to know. Still... I am not convinced of NeoSyrex's claim that the initial seed is always a 2-byte value. (0x0000 -> 0xffff) For example if you take chrish's Mewtwo example, the supposed "initial seed" is nearly 200k RNG clicks prior. That doesn't make any sense...

Thanks, I'll get started tomorrow. I'm hope that I catch up soon...

[Arseus]

Just a small update in that I have updated the first post with a couple quick notes concerning current programs and the FR/LG seed (ie. how an accurate seed cannot be obtained from RNG Reporter or PokéRNG as of yet), and I have also pasted in mingot's reverse RNG formula for those who are interested.

On another note, I know I have stated before that my activity level is rather low due to school and such, but this will especially apply for the next few weeks. I have a decent-sized research project for school, and as such, I will not be on Smogon much until it is at a satisfactory level of completion. Please do not stop posting in the thread just because I am absent; as demonstrated above, there are plenty of users who can help you out or comment on your research! I realize that this thread is not exactly overflowing with new information at this point, but I thought that I may as well post this in case I don't get on for a while.

[NeoSyrex]

Originally Posted by Fat Arseus

I had rather hoped that you had actually discovered how the games seed, but at least we have something new to work with.

Yeah, when I was trying to figure out FR/LG's RNG my primary concern was how to apply it in a practical context, in this case being how to use it to get a high-IV Zapdos for the VGC. Because I was pressed for time I didn't bother trying to figure out all of the inner workings of the game's RNG, and focused only on the parts that were necessary to know in order to abuse it. In this case, knowing the formula the game uses to create the starting seed doesn't matter that much because "calibration" needs to be done to figure out which starting seeds you commonly get anyway. Yes, with a complete formula one could figure out delays and button press combinations that result in better spreads. But like I said, it's harder in FR/LG than any other game to land on the correct starting seed due to the button input affecting it- you can't mash through the menus like you can in the other games. So it would be even more difficult trying to control these factors. But still, it is something that would be nice to know regardless.

Originally Posted by Fat Arseus

Have you been able to hit the same seed across multiple savegames? In other words, has the "seed you land on often" been the same even after saving, and doing it again? If so, then "time in-game" may not be a factor after all.

It's been a long time since I've worked with FR/LG and I unfortunately don't remember all of my findings. I wrote them down somewhere but I can't find the file at the moment. I do know I specifically tested this at one point because I too was curious if there was some data in the save file used to create the starting seed, but I can't remember what the results were. I'm leaning towards no, but I'm not 100%. I'll keep checking to see if I can find this. But I wouldn't be surprised if the only two factors were time and button presses on starting screen.

So far the only thing I was able to find was a file with some data on how button inputs affect the starting seed. But it seems so random I'm not sure if the data is accurate or not.

http://www.mediafire.com/file/tzk2zu...SeedFinder.exe

Here's the program I was talking about. Like I said it's pretty simple, just enter the seed you calculated and it will return possible starting seeds, and how many frames earlier they occured. Usually the first one (the lowest number of frames backwards) will be the correct starting seed, although there are exceptions, especially if you waited a while to calculate your seed.

[Shelcario]

Wow this sounds really interesting to manipulate the RNG in FR/LG but, I have a question and that is do we actually know how to manipulate their RNG's yet?, and How can I help in the RNG research?

[mattj]

Originally Posted by Fat Shelcario

Wow this sounds really interesting to manipulate the RNG in FR/LG but, I have a question and that is do we actually know how to manipulate their RNG's yet?, and How can I help in the RNG research?

Please Read the OP. It's got everything we know, and what we would like to see done there. Oh, BTW, I've been following for a while now and I'll be picking up a second copy of Fire Red to help out this weekend. I wan's me mah Jolly flawless Mewtwo! >:P Oh, and Syrex, that was an awesome Zappy you used at Worlds 09. It was very impressive! Glad to know there are some concrete results to your research!

[ultimatedra]

I would love to help in any way i can. Let me know if there is anything i can do. I have both FireRed an LeafGreen.

And something i noticed when i was looking online, was that there where 2 different runs of these games. The 1st run of the games came with the Wireless adapter, but on the 2nd run without the adapter, the game code was changes slightly - most likely trying to prevent cheating devices. Thoughts on this?

[chrish]

Hi Arseus,

Thanks for the post; I am embarassed to say that I didn't see that there were multiple pages that followed my last post, so I am just now absorbing the new information.

We are happy for your suggestions and we will try and make some progress implementing them in the n.f.

We will try using PokeRNG so that we can be consistent. I presume that it is the one that I found on the files pump site. I did find the RNG Reporter results puzzling, so thanks Newts Ute for the suggestion - will check it out, and run all of our MewTwo data through it and examine the results.

Since my last post the only thing new is that we verified that we could hit the same frame after a soft reset (had to be true, but wanted to see it for ourselves), however we did not hit any of the frames that we hit after a hard reset - probably due to the slightly different timing at the beginning or too small a sample space.

[Arseus]

Just dropping in to make a quick post; my activity will not markedly increase after this point, unfortunately. Conversely, I may be looking at inactivity in December as well. Regardless, here is what I have to say at the moment:

Originally Posted by Fat NeoSyrex

Yeah, when I was trying to figure out FR/LG's RNG my primary concern was how to apply it in a practical context, in this case being how to use it to get a high-IV Zapdos for the VGC. Because I was pressed for time I didn't bother trying to figure out all of the inner workings of the game's RNG, and focused only on the parts that were necessary to know in order to abuse it. In this case, knowing the formula the game uses to create the starting seed doesn't matter that much because "calibration" needs to be done to figure out which starting seeds you commonly get anyway. Yes, with a complete formula one could figure out delays and button press combinations that result in better spreads. But like I said, it's harder in FR/LG than any other game to land on the correct starting seed due to the button input affecting it- you can't mash through the menus like you can in the other games. So it would be even more difficult trying to control these factors. But still, it is something that would be nice to know regardless.

It's been a long time since I've worked with FR/LG and I unfortunately don't remember all of my findings. I wrote them down somewhere but I can't find the file at the moment. I do know I specifically tested this at one point because I too was curious if there was some data in the save file used to create the starting seed, but I can't remember what the results were. I'm leaning towards no, but I'm not 100%. I'll keep checking to see if I can find this. But I wouldn't be surprised if the only two factors were time and button presses on starting screen.

So far the only thing I was able to find was a file with some data on how button inputs affect the starting seed. But it seems so random I'm not sure if the data is accurate or not.

http://www.mediafire.com/file/tzk2zu...SeedFinder.exe

Here's the program I was talking about. Like I said it's pretty simple, just enter the seed you calculated and it will return possible starting seeds, and how many frames earlier they occured. Usually the first one (the lowest number of frames backwards) will be the correct starting seed, although there are exceptions, especially if you waited a while to calculate your seed.

Glad to hear from you again NeoSyrex. Thank you for posting your program; from experience with the last Emerald thread, I know that file-hosting sites such as MediaFire (though in the old thread it was MegaUpload, I assume these sites are very similar) may not work for all users, so I have hosted it as an attachment on the first post for the time being. And yes, I do realize that it will be much harder to abuse in FireRed/LeafGreen than in any other game we know of (this much was clear when the topic first began). However, all that can be done to elucidate the RNG mechanics should be done, and we are making progress of some sort.

As for your speculation concerning the starting seed, I am inclined to believe that it is indeed created by a smaller number of variables than I had initially thought. I do not know how or if button presses factor in (more research needs to be done in this area), but it seems clear that time spent in the title screen is one of the factors. I too am leaning towards "no" on whether something in the save file affects the seed, given the identical Mewtwo which was encountered by two of our researchers.

Originally Posted by Fat ultimatedra

And something i noticed when i was looking online, was that there where 2 different runs of these games. The 1st run of the games came with the Wireless adapter, but on the 2nd run without the adapter, the game code was changes slightly - most likely trying to prevent cheating devices. Thoughts on this?

Yes, I am aware of the two different runs of the games; my own version of FireRed is from the second run (denoted by the lack of a wireless adapter). However, I am extremely doubtful that any coding related to the RNG would have been altered. I cannot assign you a specific "job", but things that we are looking into have been outlined both in the original post and in further posts within the topic.

Originally Posted by Fat chrish

I am embarassed to say that I didn't see that there were multiple pages that followed my last post, so I am just now absorbing the new information.

We are happy for your suggestions and we will try and make some progress implementing them in the n.f.

We will try using PokeRNG so that we can be consistent. I presume that it is the one that I found on the files pump site. I did find the RNG Reporter results puzzling, so thanks Newts Ute for the suggestion - will check it out, and run all of our MewTwo data through it and examine the results.

Since my last post the only thing new is that we verified that we could hit the same frame after a soft reset (had to be true, but wanted to see it for ourselves), however we did not hit any of the frames that we hit after a hard reset - probably due to the slightly different timing at the beginning or too small a sample space.

No problem concerning your small error, chrish; we are all bound to make some from time to time. Allow me to make another suggestion: try using NeoSyrex's FR/LG Seed Finder attached to the first post (or the MediaFire version if that works) to help you find your initial seed. In any case, thank you for your research and I look forward to your further findings!

[chrish]

Thanks Arseus for your post.

We have some new findings to report:

0) We eliminated the following 10 beat delay:
Pressed A after "Wild MEWTWO appeared!" is displayed.
As expected, this made no difference.

1) We advanced the time in game (from 23:07 to 23:08) and immediately duplicated
a result from the 23:07 time in game. It seems likely that time in game makes no difference.

2) Out of 62 results we hit 7 duplicates ( ~ 1 in 9 ); from having spent hundreds
(honest!) of hours hitting spreads in Emerald for breeding/capturing, I find that
this ratio feels very similar - we have concluded from this that it is highly likely
that there is only one significant timing factor that controls the result - at least for
MewTwo. We intend to eliminate one piece at a time in an attempt to prove this.

3) Here are our 62 results in CSV format (metronome set to 120 beats/minute)
NOTE: We did not compute frames - any help with this would be appreciated.

Ref,Nature,HP,Atk,Def,SpA,SpD,Spe,reset type,beats,time in game,duplicated row #
1,Modest,19,5,7,18,22,0,hard,70/10/80/10/20,23:07,
2,Calm,26,5,1,22,28,15,hard,70/10/80/10/20,23:07,
3,Relaxed,21,7,9,7,21,22,hard,70/10/80/10/20,23:07,
4,Adamant,29,12,1,16,20,29,hard,70/10/80/10/20,23:07,
5,Serious,19,24,17,24,9,2,hard,70/10/80/10/20,23:07,
6,Naïve,28,14,3,30,17,24,hard,70/10/80/10/20,23:07,
7,Careful,29,0,26,16,4,12,hard,70/10/80/10/20,23:07,
8,Timid,5,7,18,17,26,18,hard,70/10/80/10/20,23:07,
9,Calm,18,13,19,28,7,11,hard,70/10/80/10/20,23:07,
10,Timid,12,8,29,0,31,22,hard,70/10/80/10/20,23:07,
11,Careful,29,0,26,16,4,12,hard,70/10/80/10/20,23:07,duplicate of 7
12,Jolly,6,18,26,23,10,13,hard,70/10/80/10/20,23:07,
13,Lax,2,6,4,30,12,29,hard,70/10/80/10/20,23:07,
14,Adamant,28,4,14,10,4,31,hard,70/10/80/10/20,23:07,
15,Adamant,13,23,10,27,24,28,hard,70/10/80/10/20,23:07,
16,Modest,19,5,7,18,22,0,hard,70/10/80/10/20,23:07,duplicate of 1
17,Impish,22,29,30,14,3,28,hard,70/10/80/10/20,23:07,
18,Mild,2,17,3,30,17,6,hard,70/10/80/10/20,23:07,
19,Careful,26,8,8,20,26,6,hard,70/10/80/10/20,23:07,
20,Adamant,13,30,22,4,29,23,hard,70/10/80/10/20,23:07,
21,Rash,29,22,2,14,11,7,hard,70/10/80/10/20,23:07,
22,Lax,2,6,4,30,12,29,hard,70/10/80/10/20,23:07,duplicate of 13
23,Gentle,11,16,8,11,0,28,hard,70/10/80/10/20,23:07,
24,Lonely,31,13,7,17,5,5,soft,70/10/80/10/20,23:07,
25,Bold,22,16,20,10,13,17,soft,70/10/80/10/20,23:07,
26,Modest,0,24,13,27,24,25,soft,70/10/80/10/20,23:07,
27,Rash,15,1,13,11,12,14,hard,70/10/80/10/20,23:07,
28,Jolly,14,11,12,30,22,13,hard,70/10/80/10/20,23:07,
29,Impish,26,12,19,23,7,28,soft,70/10/80/10/20,23:07,
30,Hasty,20,0,16,31,16,30,soft,70/10/80/10/20,23:07,
31,Sassy,21,8,25,17,16,31,soft,70/10/80/10/20,23:07,
32,Timid,26,0,28,12,2,12,soft,70/10/80/10/20,23:07,
33,Bold,31,8,13,2,23,25,soft,70/10/80/10/20,23:07,
34,Quirky,23,5,20,18,0,1,soft,70/10/80/10/20,23:07,
35,Relaxed,10,6,20,21,5,28,soft,70/10/80/10/20,23:07,
36,Lonely,9,14,16,0,14,15,soft,70/10/80/10/20,23:07,
37,Quiet,15,7,25,25,25,13,soft,70/10/80/10/20,23:07,
38,Docile,5,17,12,3,3,17,soft,70/10/80/10/20,23:07,
39,Docile,5,17,12,3,3,17,soft,70/10/80/10/20,23:07,duplicate of 38
40,Timid,20,7,28,24,5,26,soft,70/10/80/10/20,23:07,
41,Rash,4,22,20,19,7,14,soft,70/10/80/10/20,23:07,
42,Lonely,29,9,20,30,30,27,soft,70/10/80/10/20,23:07,
43,Brave,28,1,20,23,21,15,soft,70/10/80/10/20,23:07,
44,Brave,26,24,5,2,30,16,soft,70/10/80/10/20,23:07,
45,Bold,3,1,15,28,12,1,soft,70/10/80/10/20,23:07,
46,Quirky,17,3,3,19,16,12,soft,70/10/80/10/20,23:07,
47,Rash,21,30,17,21,11,9,soft,70/10/80/10/20,23:07,
48,Careful,19,9,29,23,5,3,soft,70/10/80/10/20,23:07,
49,Sassy,20,5,26,16,20,14,soft,70/10/80/10/20,23:07,
50,Timid,20,7,28,24,5,26,soft,70/10/80/10/xx,23:07,
51,Quirky,30,4,13,2,2,6,soft,70/10/80/10/xx,23:07,
52,Hardy,27,7,5,13,28,3,soft,70/10/80/10/xx,23:07,
53,Quiet,12,23,25,20,17,2,soft,70/10/80/10/xx,23:07,
54,Rash,21,30,17,21,11,9,soft,70/10/80/10/xx,23:07,duplicate of 47
55,Docile,4,7,19,13,13,6,soft,70/10/80/10/xx,23:07,
56,Quirky,30,26,11,22,12,19,soft,70/10/80/10/xx,23:07,
57,Jolly,18,14,17,1,31,0,soft,70/10/80/10/xx,23:07,
58,Naïve,26,26,30,7,19,4,soft,70/10/80/10/xx,23:07,
59,Naïve,26,26,30,7,19,4,soft,70/10/80/10/xx,23:07,duplicate of 58
60,Brave,26,24,5,2,30,16,soft,70/10/80/10/xx,23:08,duplicate of 44
61,Hasty,5,13,14,18,31,0,soft,70/10/80/10/xx,23:08,
62,Jolly,9,21,11,9,11,8,soft,70/10/80/10/xx,23:08

[NixHex]

I've been trying to find some info on the RNG all non-emerald gen 3 games and I'm so glad I stumbled across this. In cartridge form I have FireRed (issue #1 with wireless adapter, though probably, as stated above, is irrelevant), and LeafGreen issue #2. Right now I am starting to soft-reset for Mewtwo in LeafGreen, but I have a simple question that should put me in the right position to start helping with this.

In the NeoSyrex's SeedFinder program, what "seed" do I initially enter? Is this the PID?

As a side note, I just got my bachelor's in EE and in my final communications class we covered simple binary RNG's (no more than 8 bits, but the principles are the same as the famous ones like the LCRNG and the Mersenne Twister) which spurred my overall interest in RNG's. I do a lot of programming in MATLAB which is a really easy language to do quick work like this. Has anyone else worked with this program? The coding is quick and somewhat informal but it's an extremely powerful hunk of software. I'd like to help out with this as much as possible but I don't have a real consistent schedule. I'll read over the OP a few times and try to decide where my niche lies. Thank you guys for all of the prior information and I'm hoping I can contribute to this cause!

[chrish]

Hi NixHex,

Thanks for your post.

I am interested to know how to tell the difference between issue #1 and issue #2 cartridges.

We have 10 or so copies of LG/FR and would like to qualify our previous research by issue # (we should be able to determine whether it makes a difference as we accumulate more and more data points - we will need to play the issue # different from the one for which we are currently gathering data to the point where we are able to capture MewTwo ).

[NixHex]

Unfortunately I am only at Silph Co in my firered (issue#1 w/wireless) so i'm gonna have to start hauling ass to get to mewtwo. you know what? I can get to articuno easily as I have surf and strength (completed fuchsia before saffron), but I need to know if anyone knows whether Seafoam Islands are a quiet or noisy area.

Again, I'll be glad to help in any way that I can.

[NixHex]

Also chrish, by hard reset to you mean, switch the gba on at t=0? i'm trying to hit some of your spreads using your timing and haven't hit any so far but I'm still going at it.

[chrish]

Originally Posted by Fat NixHex

Also chrish, by hard reset to you mean, switch the gba on at t=0? i'm trying to hit some of your spreads using your timing and haven't hit any so far but I'm still going at it.

Thanks for the post; I don't know whether the seafoam islands are quiet or noisy, but the environment where Articuno resides is so similar (both at bottom of a cave) to that of MewTwo I would be willing to bet that it is the same as MewTwo.

Yes - hard reset = turning the game on with the power switch; the total elapsed time before getting to the encounter is smaller with a soft reset.

If you can tell me how to distinguish issue #1 from issue #2, I can tell you which one we are currently using for MewTwo, and I can get started playing another game (hopefully with the other issue type) through to MewTwo.

[megamonk4]

Originally Posted by Fat Shelcario

I have a few questions:

1. What are the different numbers that are displayed in the FR/LG Seed Finder when you enter a seed?

2. What is this "metronome beating" people are talking about?

3. What is the process of figuring out if an area is "noisy" or "quiet?"

4. I'm guessing that you use a sweet scenter to encounter wild pokemon easily instead of running around in the grass, water, or cave?

5. Does running around in the grass, water, or cave advance the RNG of FR/LG?

6. Would pressing "Start" and then the "A" button at a specific time each time when resetting be a viable way of finding a duplicate seed?

7. How do you calculate the initial seed with Mingot's reverse RNG formula?


For the last question what I'm stuck on is how do you mulitply and add the seed with those other weird numbers example:
0xeeb9eb65

So what do you understand? You do know that this is still the beginning of a lot of work, right?

1. I don't understand

2. As in using a metronome to use for button-pressing and other uses. Like music(which is what a metronome I used for)

3.This something that is still in process, but to figure this out, either the frame is "skipping" like 1-3-5...indicates that its a noisy area. Quiet does it by one.

I have to go for now... Someone else is like to answer or help explain.

[Shelcario]

Well thanks for the answers but, I've decided that I want nothing to do with RNG abuse and just breed the normal, fun way:D