Past Gen RNG Research

[Kaphotics]

Yeah the further down the list is what threw me off. I thought it was the first 6, but for my tests it looked like place 8 onwards iirc. This was with an English ditto and a japanese aaken.

 

[The Shadow Knight]

Here's a neat trick for those that are breeding for final results on a certain seed:

Provided you're breeding with Everstone and a Power Item you'll be likely to consistently hit the same frame with a certain Nature/Gender/Ability/Inheritance.
Now, if these traits on that frame aren't to your liking you can attach another Power Item to your parent and get a new frame that way!

So you could try all Power items and see if you hit a frame you can use. If the inheritance doesn't work out on a certain frame but the other traits are as required you can swap out a parent to one with the right IV's as long as you make sure to use the same nature.

 

[Dan Dan]

I finally found a (semi) decent spread!

29/9/7/x/7/30

Yes, not on the same level ask TSKs, but I'm sure I can find something decent to breed. Might be able to do my HP Rock Ulgamoth yet!

Question regarding NPCs on the daycare route - I believe there are four, possibly five on the route. Can their movement be predicted? IE each time you reset on a certain time and enter the game, their movements will be the same? or are they random each reset?

im so close ;-;

Timid - 31/9/7/31/31/30
Timid - 29/31/7/31/31/30

 

[Kaphotics]

Question regarding NPCs on the daycare route - I believe there are four, possibly five on the route. Can their movement be predicted? IE each time you reset on a certain time and enter the game, their movements will be the same? or are they random each reset?

They should all move in the same way if you hit the same seed. For pretty much all seeds you can get the egg or pause without any difference in frame advancement. If you are quick you can hit every possible frame by using the (saving= +1) to hit consecutive frames over a bunch of resets. When you save, the RNG seed changes, thus changing the movement of the NPCs so be wary of that.

Predicting their movement is only a guessing game like last generation.

 

[mattj]

The RMNPCs will move the exact same amount in the exact same amount of time if you hit the exact same seed UNTIL you save again. Every time you save the NPCs are in different positions. It was the same way in 4th Gen.

 

[Dan Dan]

Ok, so say I wanted to map the first 10 or so frames of this seed (29/9/7/x/7/30). Would the steps more or less be:

save in front of daycare man and hope NPCs aren't moving (lol)
reset at specific date/time, enter game and collect egg asap = frame 1
reset at specific date/time, enter game and save asap, then collect egg asap = frame 2
reset at specific date/time, enter game and save 2x asap, then collect egg asap = frame 3...

so on (providing the NPCs aren't moving in between saves..). I can see this method having problems if there are already NPCs moving on your save (you can check this by trying for your "frame 1" egg a few times, and see if the IVs change?), or if they move in between saves/before you speak to daycare man. Unfortunately you can't see ANY NPCs standing right in front of the man, and all of them are spread pretty evenly throughout the route.

 

[The Shadow Knight]

Yeah the further down the list is what threw me off. I thought it was the first 6, but for my tests it looked like place 8 onwards iirc. This was with an English ditto and a japanese aaken.

I'm getting the same results, the IV's are the numbers from frame 8 to frame 13 in RNG reporter.

 

[mattj]

Could either of you elaborate on how you're finding those numbers? I looked at two different seeds last night, but couldn't find the seeded egg IVs I was getting.

 

[Dan Dan]

I'd like to know how this all relates to finding out IVs using RNG reporter - all my data is just using Metalkid for IVs and a pen/paper writing everything I do down. I don't think I can find my actual initial seed without using an ARDSi.

To find the "base IV set", I'm just SRing on the same time/date and waiting various amounts of times before collecting the egg, and then noting which different IVs are passed - you can see what's "underneath" previous sets. eg:

all resets done on 20:17:20

egg 1 - 2/9/7/23/5/30 (A/9/7/B/B/30) so far it's x/9/7/x/x/30
egg 2 - 29/16/7/20/7/31 (29/B/7/B/7/B) now it's 29/9/7/x/7/30

different parents:

egg 3 - 31/9/7/31/31/30 (A/9/7/B/B/30)
egg 4 - 29/31/7/31/31/30 (29/B/7/B/B/30)

Still trying to finalise the SAtk :P

 

[mattj]

Ha! Sorry. What I was asking was what settings did they have the researcher on when they found the egg IVs. I had mine set up like OD had described, and it showed the monster IVs, but I didn't see the seeded egg IVs that appeared on any of the seeds I looked at.

 

[Kaphotics]

Ha! Sorry. What I was asking was what settings did they have the researcher on when they found the egg IVs. I had mine set up like OD had described, and it showed the monster IVs, but I didn't see the seeded egg IVs that appeared on any of the seeds I looked at.

IRNG, with the MTRNG value at the start. The egg base IVs are from 8-13 if you are holding just an everstone. It might be different for power items and normal/DW breeding.

 

[ΩDonut]

IRNG, with the MTRNG value at the start. The egg base IVs are from 8-13 if you are holding just an everstone. It might be different for power items and normal/DW breeding.

It shouldn't be different with Power items; they just change which IVs are passed down from the parents.

 

[The Shadow Knight]

IRNG, with the MTRNG value at the start. The egg base IVs are from 8-13 if you are holding just an everstone. It might be different for power items and normal/DW breeding.

It's also the same for non-everstone breeding.

@mattj: Are you checking your seed before collecting the egg? I also couldn't find matching IV's when checking the seed after collecting.

 

[Expert Evan]

I had some progress finally last night with the egg RNG as I was able to get something at approximately 11/12/2010 20:50:10 boot time with jolly and B/B/29/9/A/31 where father had the power item for sp.def and jolly mother with everstone. Although about 1 in 8-10 attempts with different parents but similar hold items and mother being jolly yielded me the following results even though I was hoping for different abilities:

Jolly male kibago: 31/31/29/9/31/31 ability: Rivalry
Jolly male taillow: 31/31/29/9/31/31 ability: Guts (mother had dreamworld ability of scrappy)

I tried similar with the wotter where I was using jolly ditto and ended up getting same IV spread but a different nature after like 50 re-attempts.

I tried adding the code from Kaphotics' post for IV MTRNG as per mattj's suggestion on IRC but I couldn't seem to understand how to use it as mattj had already logged off when I tried it.

 

[The Shadow Knight]

I had some progress finally last night with the egg RNG as I was able to get something at approximately 11/12/2010 20:50:10 boot time with jolly and B/B/29/9/A/31 where father had the power item for sp.def and jolly mother with everstone. Although about 1 in 8-10 attempts with different parents but similar hold items and mother being jolly yielded me the following results even though I was hoping for different abilities:

Jolly male kibago: 31/31/29/9/31/31 ability: Rivalry
Jolly male taillow: 31/31/29/9/31/31 ability: Guts (mother had dreamworld ability of scrappy)

I tried similar with the wotter where I was using jolly ditto and ended up getting same IV spread but a different nature after like 50 re-attempts.

I tried adding the code from Kaphotics' post for IV MTRNG as per mattj's suggestion on IRC but I couldn't seem to understand how to use it as mattj had already logged off when I tried it.

To use the RNG code you need to do the following:

Go to your Trainer card's badges screen, press Select. Now go back to your main trainer card and take note of your money and your trainer ID(on the backside of your trainer card), they will be in decimal notation.
Let's say your amount of money converted to Hexadecimal is XXXX, and your TID converted to Hex is YYYY.
Your seed is, XXXXYYYY.

 

[ΩDonut]

This is something I should have derived weeks ago with X-Act's technique, but couldn't figure out what I was doing wrong. Only this morning I realized I made a simple arithmetic error.

Here is the reverse RNG formula:

x[n + 1] = x[n] * 0xDEDCEDAE9638806D + 0x9B1AE6E9A384E6F9

As the name implies, it gives the results of the RNG in reverse order, which will be very helpful in finding the initial seed of the PID RNG, with the program I mentioned in this post.

 

[The Shadow Knight]

I have noticed that when SR'ing for egg IV's on a certain date and time (with precision to the second)you'll get quite a few different seeds. There's usually one seed that's most frequent and a couple of other ones. This RNG's supposed to not have any start-up delay taken into count, right? What could be the cause for this?

 

[Dan Dan]

I have noticed that when SR'ing for egg IV's on a certain date and time (with precision to the second)you'll get quite a few different seeds. There's usually one seed that's most frequent and a couple of other ones. This RNG's supposed to not have any start-up delay taken into count, right? What could be the cause for this?

I can somewhat back this up with resets, but wouldn't know what is causing it. Perhaps the shaking grass/fishing spot encounters could have a bearing on the frame advancing as well?

parents:
A - 31/31/25/31/31/31 Everstone (Timid)
B - 31/31/31/8/31/31 Power Item (Timid)

seed:
C - 29/9/7/3/7/30

"ideal":

29/9/31/31/31/30 pass these IVs

but no matter what I do, I get stuff like this:

31/9/7/31/31/30
29/31/7/31/31/30
29/31/7/31/31/30
17/22/9/31/31/17 wut
31/9/7/31/31/30 (save 1x then collect)
31/23/25/31/30/31 wut
31/22/25/28/0/31 (wait 15 seconds)
31/9/31/3/31/30 spam a
17/31/31/28/0/31 wut
17/22/31/31/0/31 waiting indeterminable amount
29/31/31/3/7/31 (saving 1x then collect)
17/22/31/31/0/31 (saving 2x then collect)
29/31/31/3/7/31 (saving 3x then collect)

and i gave up there. at first i thought the "other" base iv set i was getting (17/22/?/28/0/?) was from me starting on the wrong second, but I was very precise with my timings and noted everything down. I dont think continually saving the game is a very good way of frame advancing.

 

[Kaphotics]

I have noticed that when SR'ing for egg IV's on a certain date and time (with precision to the second)you'll get quite a few different seeds. There's usually one seed that's most frequent and a couple of other ones. This RNG's supposed to not have any start-up delay taken into count, right? What could be the cause for this?

See mattj's posts, he showed that there can be some other seeds you may hit. How, I'm not sure because I haven't witnessed it myself.

Awesome OD!

 

[mattj]

I have noticed that when SR'ing for egg IV's on a certain date and time (with precision to the second)you'll get quite a few different seeds. There's usually one seed that's most frequent and a couple of other ones. This RNG's supposed to not have any start-up delay taken into count, right? What could be the cause for this?

You and me both brother. :/

Using the precision of emloop, I SR'd on a bunch of different times. I found a very clear pattern that around 2-3 seeds (possibly more in some cases) were being repeated ONLY within 1 second of each other. Never 2 or 3 or anything like that. 1 second "seemed" to be yielding 1 common and 2-3 less common seeds.

 

[Dan Dan]

Yay I got an ARDSi! Finally got the MTRNG code to work after updating the firmware and eventually worked out how to find my initial seed, and get the predicted IVs show up in researcher. I'm don't think anyone else has researched consecutive eggs, and where they show up yet, but here are the first 5:

1st egg - row 8-13
2nd egg - row 25-30
3rd egg - row 38-43
4th egg - row 63-68
5th egg - row 84-89

*ignore, doesn't work >_>*

Hide

Parent A - 7 / 13 / 19 / 18 / 8 / 14
Parent B - 3 / 2 / 11 / 27 / 25 / 10

Reset 1: 14-11-2010 (14th Nov) 19:26:39
seed: 20AC3117
egg 1 - B/A/23/28/A/17 (predicted in reporter rows 8-13: 27/25/23/28/24/17)
egg 2 - 29/B/B/A/22/12 (predicted in reporter rows 25-30: 29/20/31/22/22/12)
egg 3 - A/12/1/A/A/13 (predicted in reporter rows 38-43: 22/12/1/2/9/13)
egg 4 - A/A/14/14/A/6 (predicted in reporter rows 63-68: 4/21/14/14/28/6)
egg 5 - 12/A/B/B/22/18 (predicted in reporter rows 84-89: 12/28/29/29/22/18)

Reset 2: 14-11-2010 (14th Nov) 20:07:40
seed: 039B90C7
egg 1 - 29/15/A/10/B/B (predicted in reporter rows 8-13: 29/15/8/10/24/17)
egg 2 - 0/18/B/26/A/A (predicted in reporter rows 25-30: 0/18/25/26/0/19)
egg 3 - A/8/B/B/18/29 (predicted in reporter rows 38-43: 24/10/30/21/10/6) doesnt match, actual location rows 53-58 (20/8/15/25/18/29)
egg 4 - A/11/2/A/26/A (predicted in reporter rows 63-68: 22/20/30/0/30/11) doesnt match, actual location rows 75-80
egg 5 - A/5/B/31/B/15 (predicted in reporter rows 84-89: 26/6/4/31/10/7) doesn't match, actual location rows 100-105

edit: added my actual data, and going to do another reset and see if they remain constant.
edit2: it doesn't match past the 2nd egg.. i'll do another reset, but I think I'll get similar results.

 

[ΩDonut]

Yay I got an ARDSi! Finally got the MTRNG code to work after updating the firmware and eventually worked out how to find my initial seed, and get the predicted IVs show up in researcher. I'm don't think anyone else has researched consecutive eggs, and where they show up yet, but here are the first 5:

1st egg - row 8-13
2nd egg - row 25-30
3rd egg - row 38-43
4th egg - row 63-68
5th egg - row 84-89

Nice find! I'll look into what makes the IV RNG jump ahead to different rows. One question: were you using any Power items?

Anyway, I've finished the program that allows cartridge users to find the value of their initial seed WITHOUT needing an emulator or AR. It's attached to this post, along with the source code (C#) so you can compile it yourself if you like.

And the instructions:

[1] Start the game at the DS date\time that gives you the desired initial seed.
[2] Select "New Game".
[3] Begin the process of starting a new game, up to the point you can open the menu. (Obviously, do not save at any point.)
[4] When Professor Araragi asks you a Yes\No question right after you input your nickname, choose YES. (Choosing NO even once at this point means you will have to start again from step 1.)
[5] Look at your new Trainer Card and record the trainer ID.
[6] Repeat steps 1-5, but when Araragi asks you Yes\No after the nickname, say NO once to go back to the nickname screen, then say YES. Record the trainer ID that results.
[7] Repeat step 6, but this time say NO twice before saying YES. Record the trainer ID that results.
[8] Repeat step 7, but this time say NO three times before saying YES. Record the trainer ID that results.
[9] Enter the four IDs into this program, in the order you obtained them.
[10] Start the game at the DS date\time that gives you the desired initial seed. (Be sure you saved in an area with wild Pokemon, preferably a cave.)
[11] Select "Continue". Catch the first Pokemon you encounter and record its IVs. Input these IVs into the program.

The program then gives you a number you can use to find your real initial seed. Why is it not the real initial seed? Well, this should explain it:

Hide

57E9E104 20C21C61 <-- the VERY initial seed, the first 32 bits (57E9E104) becomes the seed for the IV RNG
0DF5A2B5 8092BA08
3AD67B10 B5D04BEB
B4D1E454 6B62557A
FBC1A5A7 CC06A1E5
36F3BE37 14AB0B1C
9352D533 CE57FCCF
6C61BC97 BCD1236E
6CDAB4F9 CAA17729
C9F302D9 E8BE92F0 <-- value of the RNG before selecting Continue or New Game
BB04C96A 14890773
72E0FC1A 27581A22
58A07938 BAD3202D
D7C97D76 8EE56584
31295A6C 9A974FD7
83D91838 DA742D96
467A65EC E37BE0F1
EB231190 DDFE56D8
22995EC4 31E079FB
C3B4DBB9 08DB11CA
024311EE 6F9FBD75 <-- [b]value of the RNG when first naming your character, this is the number the program gives you[/b]
44C76443 10C2FAEC <-- say NO to Araragi right after naming your character and the RNG advances here
58854CBA 61CFE9DF <-- and so on
DFAEC135 546F3ABE
79835702 99AA79B9
F1395043 3B9BA5C0
27F4955B 5882C383
FE0AA925 82F3DC72
A6172386 4A7399BD
2F45F669 2D336B54
5CDA6282 B92BEAE7
E9D9841B 6F80EAE6
24C5CBDC 70026181
CBA3FA01 8DCC1FA8
46D22AB6 EC38040B
BF826728 ECB71A1A
A7FCBC1A 1011D505
BB48E135 D15256BC
28426944 553172EF
EC6A67AF 0BB3DE0E
9888FE71 DA54B849
7D045E9F 50316490
CC237B5B 92645B93
A91FF5D7 FBC56AC2
5DD633EC BE798F4D
555D047E 9AE75D24
D324A9AB AA42A1F7
BDEC9671 76DEB436
E860B2CC C9EE9E11
746FEAC6 DC591478
13B3E2CA CA87EA1B
C757E686 6ECB6E6A
561CDC4B 5165E895
BE6319F7 DEE61E8C <-- value of the RNG after reloading the game

As the diagram shows, the RNG when reloading the game is many frames ahead of the RNG frame when the trainer ID is generated.

So, you would have to plug the seed the program gives you into another program (that I will be releasing shortly in another post) that spits out the natures, gender, shininess, etc. for each RNG result. You would then compare this list to the natures of Pokemon you catch, but I'll post full details then.

Mind you, although the source is fairly readable it's not nearly as organized as I would like it to be. I'm not going to clean it up, though - the program will find your seed if you got all the numbers right, and won't if you didn't. My main focus is to find out how the game seeds the RNG, and I paused work on that so I could make this program for people to work with in the meantime. For that reason this technique will not be programmed into RNG Reporter.

 

[Sniffles]

I have come up with a theory that if mattj is right about the breeding thing with the running pokemon, (if they stop in the same exact spot, the egg's IVs will be the same), could we just use this with provided power items to produce perfect pokemon granted the parent's IVs are good? If the moving Pokemon advance the RNG, we could use them to advance the RNG to our desired spot. Of course, if I understand correctly you would need your initial seed to determine what you'll get. If we can come up with something to predict how much you need the RNG to advance, it would work in theory. For now, we could just use soft-reseting. Am I correct? I would test this on my cart, but unfortunately school and homework will most likely get in the way.

 

[mattj]

Haha! That's actually the general principle we're all following right now.

Reset till you find good base Egg IVs.
Equip Power items and see if you get any good PID/IV frames.
Sort through the frames by waiting different amounts of times to accept the egg (using the two spinning mons as a guage).

 

[Dan Dan]

That's pretty much what I've been doing for the last week :P