RNG Research - Page 16

Bond697 · Oct 24th, 2010, 11:53:27 PM

nevermind, my mistake.

Kaphotics · Oct 25th, 2010, 1:41:10 PM

code discussion, ignore this I made a code

Made an IVMTRNG Seed Check code, tested on DeSmuME and on a cart.
White

Code:

94000130 FFFB0000
DA000000 022151D4
D7000000 022696A4
A226969A 00000000
1226969A 00000000
DA000000 022151D4
D6000000 0226969A
D2000000 00000000

Black

Code:

94000130 FFFB0000
DA000000 022151B4
D7000000 02269684
A226967A 00000000
1226967A 00000000
DA000000 022151B4
D6000000 0226967A
D2000000 00000000

Activate with select once you are viewing your badges, or else the game will freeze or the code will not work. Once activated click your trainer card to view, no need to hold down the activator. It has no negative impacts on your game! Just exit the trainer case and everything is reverted; this code just changes what is displayed, not stored.

Money (In Hex): XXXX
TID (In Hex): YYYY

MTRNG Value: XXXXYYYY

It's pretty easy to adapt the code for other seeds, just change the bolded offsets to any other offset to check its value.

memory check template

mattj · Oct 25th, 2010, 10:15:05 PM

I like that idea ^^^

Repeated my Sweet Scent tests on Cobalon, plus I tried saving to see how that affected the PID like Kaphotics suggested. Got some results that I expected, and a few that surprised me.

...

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

10-25-10 Monday (mah 26st berf dey! :3)

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Stationary Legendaries: Do all the findings from Sweet Scented Mons apply to them too?

Pikatimer:
CD: 600 WD: 2370 CS: 1 WS: 1 FT: 30:50 ST: 30:50

Synch Date/Time: 10-17-10 18:00:00

Saved in the Fukiyose Grotto off of Rt 6 right in front of Cobalon. On foot. No Repels in use. Turning C Gear off at Continue.
Gonna see if all the old findings found while Sweet Scenting apply to Stationary Legendaries like Cobalon.
Also need to see if Saving your game advances/changes your PID just like Walking/Biking/Turning in grass all do.

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Immediately encounter

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Reset #1)
Cobalon (638)
Modest
8 / 30 / 5 / 25 / 13 / 1

Reset #2)
Cobalon (638)
Rash
15 / 9 / 30 / 9 / 30 / 25

Reset #3)
Cobalon (638)
Modest
8 / 30 / 5 / 25 / 13 / 1

Reset #4)
Cobalon (638)
Modest
8 / 30 / 5 / 25 / 13 / 1

Wonderful. That means any IVs I find for this will apply to Sweet Scenting too!
Now to verify doublecheck for RMNPCs

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Pause 30.5 seconds from final Yes after Continue, then encounter

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Reset #1)
Cobalon (638)
Modest
8 / 30 / 5 / 25 / 13 / 1

Reset #2)
Cobalon (638)
Modest
8 / 30 / 5 / 25 / 13 / 1

What is it, a 1/25 chance of randomly getting the same nature? Good enough for me.

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Turn in Place/Walk 6 times, then encounter

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Reset #1)
Up, Right, Up, Right, Up, Right
Cobalon (638)
Jolly
8 / 30 / 5 / 25 / 13 / 1

Reset #2)
Turn Around in Place, Step Left, Step Left, Turn Around in Place, Step Right, Step Right
Cobalon (638)
Jolly
8 / 30 / 5 / 25 / 13 / 1

Proof enough for me. Now to see what saving your game after hitting your seed does.

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Save Game 6 times, then encounter

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Reset #1)
Cobalon (638)
Lonely
8 / 30 / 5 / 25 / 13 / 1

Odd. It changed the PID, but differently than Walking/Turning/Biking. Maybe a different amount of frames? Or re-seeding? I remember that in HG/SS Walking/Turning gave inconsitent advancements in inconsistent incriments from seed to seed (consistent as long as you hit the same seed though). Maybe it's like that again? Don't recall biking though.

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Hearing your pokemon's cries, looking at the stats page,
and switching between the front and back views on Summary Screen, then encounter

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXX

Reset #1)
Cobalon (638)
Bashful
16 / 20 / 0 / 22 / 14 / 15

Reset #2)
Cobalon (638)
Modest
8 / 30 / 5 / 25 / 13 / 1

The thing I expected is that Turning in Place/Walking/Biking all changed the PID for Stationary legendaries in the same way that it did for Sweet Scented mons.

I honestly didn't expect to get the same IVs as when I was Sweet Scenting though. I guess it's just being used to 4th Gen RNG, where SSing advanced your frame (which included IVs), but no. Same IVs, which is nice. Find one good IV spread and it works so far for the 3 Beasts and Sweet Scented mons. Do you think it would also work for:
Victini, Reshiram, Zekrom, Kyurem, Landlos (all apparently stationary like the Beasts)
Tornelos and Voltolos (roaming)
Hihidaruma, Musharna, Zoroark, Ulgamoth, (all interactable)
Yanappu, Baoppu, Hiyappu, Zorua, Meraruba (egg), Magikarp (all gift)
Purotooga, Aaken,Omanyte, Kabuto, Aerodactyl, Lileep, Anorith, Cranidos, Shieldon (all fossil)

Or do you think they would be on different IV frames or something? Is this worth me looking into? I could use the same timer settings and therefore the same seed and see if I get the same IV frame/etc. I'm starting to not really care about the PID because it seems you can just cycle through them and pick whatever you like.

Another thing that surprised me is that Saving my game 6 times did change my PID, but it didn't change it to the same thing as when I Walked/Biked/Turned 6 times. Any idea why that might be?

Oh, and just for lols, I screwed around on the summary screen hoping it would change PIDs or IVs but no.

[edit]
Testing it now...

Kaphotics · Oct 25th, 2010, 10:32:00 PM

You could probably have 100% accurate tests now with the seed check code changed to check the main RNG. Hmm, it could probably even check the PID of the Pokemon you just caught as well! (goes back to the offsets for party pkm)

From my observations of reading the output from researcher, saving increased it by exactly 1, every time.

With a different current seed, NPCs behaved differently (fun fact).

mattj · Oct 25th, 2010, 10:35:17 PM

Quote:

Originally Posted by Fat Kaphotics

You could probably have 100% accurate tests now with the seed check code changed to check the main RNG.

From my observations of reading the output from researcher, saving increased it by exactly 1, every time.

With a different current seed, NPCs behaved differently (fun fact).

Oh yeah. I've noticed different NPC behavior based on different seeds when Wondercard and Egg abusing in Gen 4. Sometimes they'd be lazy other times they'd move like crazy. You could sometimes tell what seed you hit by their movement.

It could very well be that Saving increases it by 1, and that Walking/Biking/Turning increases it by inconsistent intervals, but consistent "overall" on the same seed, like I experienced back in HG/SS.

Oh, and what do you mean "with the seed check code changed to check the main RNG"? I thought that code did show the main RNG? Do I misunderstand you?

Bond697 · Oct 25th, 2010, 10:42:24 PM

Quote:

Originally Posted by Fat mattj

Oh yeah. I've noticed different NPC behavior based on different seeds when Wondercard and Egg abusing in Gen 4. Sometimes they'd be lazy other times they'd move like crazy. You could sometimes tell what seed you hit by their movement.

It could very well be that Saving increases it by 1, and that Walking/Biking/Turning increases it by inconsistent intervals, but consistent "overall" on the same seed, like I experienced back in HG/SS.

Oh, and what do you mean "with the seed check code changed to check the main RNG"? I thought that code did show the main RNG? Do I misunderstand you?

no, the code checks the MTRNG(IV RNG), not the main RNG. you could adjust it to check the main one, though, without too much difficulty.

basically, you take the seed you get from the code and put it in the research part of rng reporter under irng and follow omega donut's instructions(shift by 27, etc) and there's your IVs. the MTRNG seed that determines the next encounter's IVs should be the one immediately before the battle starts. at least, that's what i've found.

Kaphotics · Oct 25th, 2010, 10:43:35 PM

Agreed :)

Agreed again :)

That code will just check the value at 022151D4 (lol sniped by bond, thx bro)

If you went into the template and placed say, 0223EBFC in the code, you'd be able to check the PID of the egg you just received.
Use 022160A4 and check it once for a value, and again with 022160A8, you'd check the current RNG Seed for PIDs.

What I meant by main RNG is that you could check your initial seed to see if you are hitting the same seed every time to remove erroneous attempts, as well as checking advancement from the initial.

----------

Plugged in some offsets to make moar codes. Included a 1 button activator key if they need to be changed.

IV MTRNG Check Code [SELECT]
SEED1 Check Code [L]
SEED2 Check Code [R]
Party 1 PID Check Code [START]

codes

------

IV's for eggs aren't the same as wild Pokemon >.>

mattj · Oct 25th, 2010, 10:56:36 PM

Quote:

Originally Posted by Fat Kaphotics

e: I'll update this post with a bunch of [hidden] check locations. All thad' have to be done is just change the activator if you want to have multiple check codes.

Awesome. Start, Select, L, R maybe?

[edit]
Also, this really could just be human error, but if you look at the Resets I posted from earlier tonight, I was pretty spot on consistent. Only missed a few times, but still got IVs that I've gotten from errors in the past. However, after using this code on 5 Resets, I've gotten new IVs I haven't seen before:

...

It could just be error, but it seems odd to me. I'll keep trying...

**ΩDonut** · Oct 25th, 2010, 11:35:00 PM

Quote:

Originally Posted by Fat mattj

Original / Code
$: 30563 / 13350
ID: 16440 / 34282
23 / 12 / 18 / 5 / 2 / 15
(3 times)

Original / Code
$: 30563 / 21429
ID: 16440 / 02982
11 / 11 / 1 / 1 / 27 / 5
(2 times)

The first one is seed 342685EA, which is supposed to produce completely different IVs (12\26\12\11\10\14). The second one (53B50BA6) is supposed to produce 11\26\16\15\29\27. Something's up with the code.

Kaphotics · Oct 25th, 2010, 11:36:43 PM

Yeah Bond now says he's getting confusing results. I'll look into it, it worked fine when I was testing...

It works perfectly fine for reading the value on the emulator, and Bond initially said it worked fine on his cart for predicting (he's afk now). I tracked changing memory spots and the card $ and ID are the only ones that are changed relating to activating the code.

It works for the party 1 PID check, and the seed checks. I don't see what could be going wrong with the code other than using the wrong version...

Did it work at all for predicting IVs? Could just be ingame time/lighting related? Was the game saved very close to the time it was reloaded and checked?

Forgot to remove 20 for black for 022151D4->022151B4, FML. That's why mattj's seeds were wrong D:<
It's fixed now.

I checked to see if the IVs were predicted without the code at 022151B4, and for one test it did. The IVs were the same whether or not I used the code.

Bond697 · Oct 26th, 2010, 12:19:08 AM

Quote:

Originally Posted by Fat OmegaDonut

The first one is seed 342685EA, which is supposed to produce completely different IVs (12\26\12\11\10\14). The second one (53B50BA6) is supposed to produce 11\26\16\15\29\27. Something's up with the code.

i can't seem to get this method of predicting IVs to work at all anymore. i check the mtrng right before a battle with desmume, check in rng reporter and i have 1 set of IVs. i catch the poke and i have a completely different IV set. this is in desmume or with the code. not sure what's up, maybe different methods or something? i'm in the grass on rt3 just running around. i've thoroughly double checked the right shift, and all the other setup, but nothing is set up wrong. the only times it worked right that i can remember was when i was at the end of the game in a cave checking IVs.

e: i can verify that the code works perfectly in white. i just checked against a ram watch at 22151d4. if you take the money value and convert to hex then take the TID value and convert to hex then put the money hex in front of the TID hex, you have the right seed.

**ΩDonut** · Oct 26th, 2010, 8:30:58 AM

mattj, I'd like you to try something once you get the revised code inputted.

Can you start from October 17, 2010, 18:00:15, and turn on the C-Gear once you enter the game? Not when it gives you the option to turn it on at the start, but in-game with the C-Gear's power button. Record the time you pick yes when it asks you "do you want to turn on the C-Gear?"

You should use a timer to countdown when to pick "yes". Give your best guesstimate as to what 2000 delay might be (it takes about 1300-1400 delay to turn on the C-Gear in-game). You know how you add 200 delay* to your HGSS delay in PikaTimer\emloop\whatever to account for the gap between the moment you start the game from the DS menu and the moment it actually starts counting delay? Just use that number.

When all that is done, record the RNG result. I have a hunch, but I need to see it tested on a cart.

* number pulled completely out of my ass

mattj · Oct 26th, 2010, 9:41:11 AM

I'm actually off work today! Got some errands to run in the morning though. Just so I understand correctly, this delay for the CGear is going to be:

Quote:

from the moment I select my game on the DS Startup Screen to the moment I say "yes I want to turn my CGear on" ingame

correct?

It's just a bit confusing because I'm so used to:

Quote:

from the moment I Reset my game to the moment I press A at the Blue Continue Screen

from Gen 4.

**ΩDonut** · Oct 26th, 2010, 11:20:22 AM

Correct. It's not much different.

mattj · Oct 26th, 2010, 4:22:54 PM

Did 5 Resets under those conditions. Not sure what the results mean, but here they are:

...

It all felt kind of sloppy to me. I'm having a bit of a hard time adjusting to the new target, and I'm not even sure I know what it is for sure. When you press the CGear power button, a "Warning Message" or something pops up. There's a Yes button and No button in the top right hand corner. But when you press the Yes button, you just stay on that "Warning Message" or whatever, so I'm not sure if the CGear has really started up you know? I figured I'd hedge my bets and hit Yes when the 2nd timer ran out and just spam-tap the screen to get off that warning message as quickly as possible, but I got some sqoogly results it seems. :/

Let me know if something I did was off you know?

To make sure I understand this correctly, I'm gonna try some of my old tests but with the new IVRNG Code and putting it into the Researcher (finally upgraded my Reporter from 7.0 to 8.whatever).

[edit]
Akkk more wonky results:
wait... not wonky... i did a typo x_x

...

**ΩDonut** · Oct 26th, 2010, 7:00:19 PM

Quote:

Originally Posted by Fat mattj

The moment the Second Timer runs out, tap "Yes", then QUICKLY tap the screen to exit the "Warning Message" and actually turn on the CGear (18:00:48:33).
VERY QUICKLY check and record IVMTRNG.
VERY QUICKLY encounter Cobalon, capture, check and record IVs (less than 1 minute from CGear "Yes").

Quote:

It all felt kind of sloppy to me. I'm having a bit of a hard time adjusting to the new target, and I'm not even sure I know what it is for sure. When you press the CGear power button, a "Warning Message" or something pops up. There's a Yes button and No button in the top right hand corner. But when you press the Yes button, you just stay on that "Warning Message" or whatever, so I'm not sure if the CGear has really started up you know?

This you don't need to do very quickly. The new MTRNG seed is set the moment you click "yes", and you can wait as long as you want afterwards and get the same IVs. Don't worry about trying to click to screen to clear it as fast as possible.

EDIT: Okay, figured out what was wrong. Apparently the value stored in 022151D4 is only the IVMTRNG seed if you seed without the C-Gear. But if you do reseed with the C-Gear, it doesn't match at all.

After reseeding with the C-Gear, the value in 022151D4 was 72993771 (this was around 4160 delay, I didn't feel a need to be precise). But the IVs I got with RNG Reporter Researcher were completely different from the IVs of the critter I caught.

Now to figure out what's right.

Bond697 · Oct 26th, 2010, 9:05:56 PM

i think i've had the cgear on most of the time i've been testing. that explains so much.

mattj · Oct 27th, 2010, 11:43:19 AM

I want to see if the one version exclusive roamer we get in our games gets its IVs generated the same exact way as sweet scented, randomly encountered, and stationary legendary mons. However, they still haven't fixed the problem with backing up save files for B/W with my ARDSi, and I really don't want to risk accidentally saving after his info is generated. Do any of you guys know where's a good spot to save at when looking at the roamer?

Kaphotics · Oct 27th, 2010, 12:13:51 PM

It's pretty hard for that one since there is the ambient crazy rain (which advances the PID RNG a lot). It appears as this rain won't advance the IV MTRNG. Once you step outside of the house when getting pulled in, you follow a set path where he is generated.

Save right before you step outside and you should be able to soft reset em. (if this is what you are asking), or if you want to save him for later just save near the house but don't go near it until you do your soft resets.

safe pre-trigger spot

You should probably check the mtrng in the crazy rain at different points in time to see if it is advancing, before starting the event.

mattj · Oct 27th, 2010, 12:49:36 PM

Excellent! Just the info I needed. Yeah I just want to see if his IVs are generated on the same frame or whatever as SS'd, wild, and Stationary mons have all been so far. I don't really care about PID right now because you can pretty much SR for that once you've found good IVs. It's just that some roamers in gen 4 got their IVs from like frame 5 or whatever. I'll check to see what the rain does too, when I get off later tonight. Thanks for the help.

Kaphotics · Oct 27th, 2010, 6:21:41 PM

Been looking at the memory region for the MTRNG and how it advances, without the CGear ever being on.

I started with MTRNG Seed A, zeroed it out, and let it advance. I compared it against a nonmodified. Both were (modified) from the same state around 0.4 seconds before the first change in the value. Tested for different times, so it didn't look like it considered time, or it wasn't updated quick enough to make a difference.

Tested on white, inside a cave at victory road, no moving NPCs.

Initiating a battle (IVs generated from first line)

Code:

Viewing: 022151D4
Unaltered / Altered
D2A1C834  00000000 -- Start
F632D07C  B632D07C -- Change 1

looking around for other differences...
Viewing: 022158EC
Unaltered / Altered
11C08278  11C08278 -- Start
BAAA25C7  FAAA25C7 -- Change 1
5CCAEDD2  5CCAEDD2 -- Change 2

now back to the MTRNG...
Unaltered / Altered
6A0CFC24  6A0CFC24 -- Change 2
wat, no reliance on the first value?

Checked the rest of the memory locations for any difference after the battle screen starts its shifting (2nd change)
022152B8 [BE38DDA3 -> FE38DDA3]
022158E8 [313D1A0C -> 113D1A0C]***right next to the last change
022159D0 [5D619E07 -> 1D619E07]
02215B8C [41543FAC -> 51543FAC]

02215B94 [00000015 -> 00000012] This is the frame counter for the battle, advances 2->270 (622 decimal), with even intervals, with 00000000 not existing -> (270/2-1)*16=311 changes which then updates the MTRNG position upon looping. Starts changing once the screen fades to the appearing pokemon from darkness, not when the music starts.

02215E08 [05F23C00 -> 05F93C00] Another advancing thing that I have no clue what it does.
02215F20 [00000000 -> 00B300EC] Rapidly changing back and forth between two values, 
All 3 values have approximately the same rate.

I separated out the last 3 lines as they are not with the block the MTRNG occupies (has 0000000 in between. The last 2 don't look related at all.

So in conclusion at a quick glance it looks like MTRNG is completely reseeded in the battle. Not sure if any of the other secondary changes loop back to affect the MTRNG. I'll let it keep going and compare the 022151D4 values against eachother later. 022151D4 could just be a displayed value that is calculated from some other memory....

Does the MTRNG predict IVs for multiple pokemon in a sequence? If so it's most likely a displayed value or reseeded.

The MTRNG position advances every 311/60(fps=delay) =~ 5.2s seconds while in battle the moment the battle UI appears. The 270h counter advances at the same rate as the delay location at 023FFC3C does. I guess this information can be useful.

mattj · Oct 27th, 2010, 8:55:31 PM

Do you think that the in battle RNG advancements can be feasibly controlled?

Did a couple tests on different "kinds" of mons to see if they all got their IVs the same way, or on the same frame or whatever. Got one surprising result.

...

From the ones I tested:
Sweet Scenters, Wild Encounters, Stationary Legendaries, Overworld, and Fossil Mons all got their IVs "on the same frame", or "in the same way". I kind of expected that, and that's pretty good news.

The one that surprised me was the Meraruba Egg (636) from Red Man in house on Rt 18. I hit the same old seeds I had been hitting the whole test through but... the IVs were... "shifted" from what I expected them to be. It like, took off the last IV (Spd), stuck a new random IV in the first slot (HP), and then shifted them all to the right 1 slot. Was Riolu this way in D/P/Pt? Or Togepi in HG/SS?

Tomorrow evening I'll probably have time to test Tornelos, but I spent an hour searching for him and he kept staying in the like 4 routes I can't get to at this point in the game, so I gave up x_x (maybe I"ll use WTW next time...)

Kaphotics · Oct 29th, 2010, 2:59:54 AM

How sure are we that 023FFC3C is the Delay counter? 020A9F60 also advances at a counter like rate, but not the same rate as 023FFC3C. Because the memory location right after 020A9F60 is =[022160A4] (lol we know what this position is!)

kinda bored, so I just was looking at how the IVMTRNG was updated and from where in Cheat Engine
The 00401000 region (1FD000 long) looks pretty interesting. I can't find it in RAM search (not viewable?).
00429645 is used to write to 022151D4 initially, and 00432383 along with 00429645 is used to write to 022151D4 when the CGear is on.

004329D9, 00432383, 004288ED, 004297AC access 022151D4 when encountering a Pokemon.
All of these are also reading 022151D4 when 022151D4 advances during battle.
All 4 of these locations seem to "call 004d74d0" with 004d74d0 => sub esp, 14 (83, EC, 14) [subtract 14] along with a few other things.... that's it for now, just scribbling some stuff down.

mattj · Oct 29th, 2010, 8:56:32 PM

Finally checked Tornelos. He's another "odd" one.

...

It's like, they took the expected IVs, removed the HP, shifted Atk, Def, and SAtk to the right 1, inserted a new Random SAtk, then left SDef and Spd the same.

Is this something familiar? It's unfortunate, because you'll pretty much have to Reset for him by himself apparently. So far, all Stationary, Wild, Sweet Scented, and Fossil mons share the exact same starting IVs, but this one "sort of does" but they're shifted around, 1 is missing, and one is added.
@_@

...

actually...odd...

22 is the 7th value in the Researcher for that IV RNG seed..., the game took out the 1st value, and fit the 7th in in the middle...
...and actually...
looking back at those 3 odd results from that gift egg... they follow sort of the same pattern (drop the first, add the 7th) albeit in a different place...

does any of this make sense?

[edit]
maybe it has something to do with the fact I saved outside and had to walk through the door... I seem to remember that in HG/SS walking in or out of Kyrouduaza's room advanced your frame by 1. Maybe it's doing something odd... I can't yet save inside because they still haven't fixed the Save File Backup to where it works with B/W on ARDSi, but I can go try it on Cobalon real quick...

[edit]
or it could have been the crazy rain... I haven't Sweet Scented in the desert with that sand yet...

Kaphotics · Oct 30th, 2010, 1:20:03 AM

When I was doing the set up picture for ya, I didn't see the IV MTRNG location advance at all even throughout the entire process. I'm pretty sure the 022160[A]4(8) location was flying with the rain.

Interesting... I'll try this out for the starters to see if they get the predicted IVs.

Code:

Frame	SEED1		SEED2
0	3D072258	5C841A77	Starting Seed
1	F5C0437B	F300431A	(Nature?)
2	957B114A	73342467	Ending Seed for all 3 starters (F301431A)
3	86FB0AF5	AC53CDB7	Start Belle Battle
4	42C20E6C	385453CA	Start Cheren Battle
5	25621B5F	DE525286	They left (1)
6	141A423E	8E4E23A4
7	F3A7EF39	A3824B9D
8	96FC8140	31B933B6
9	3BB0DD03	863B8385
10	82496BF2	418042C4
11	BD80B73D	7F700746
12	64A98ED4	1337A036
13	CC476C67	91F60FDC
14	3C288266	E83AD4C1
15	02B4A701	0D93E9CB
16	BBD30B28	7A759954
17	397A6D8B	3F142134
18	D4E0399A	E6F2F5C3	They left (2), end (saving now)
19	86A6C285	F437855F	after saving
20	CBAD8A3C	DCD3BBD7
21	523C446F	80297DDF
22	6596058E	0DFF1D6D
23	6C31CDC9	094383FE
24	C36C6010	375A86C1

The PID is set the second you dismiss the Received Pokemon message, before they start talking to you again. The IV MTRNG changes when you receive your starter.

There is no passive advancement for either seed, so hitting a seed and saving to advance is the only way to go for starters.

IVs were predicted by the IV MTRNG Seed.

The frame advances by 1 when the battle screen starts flashing for each rival. The seed does not change when the MTRNG seed does during the battle.

Why does it advance so much after not advancing by much at all? Probably setting your rival's Pokemon, as they were different even with the same seeds, even between save states (thus based off of some other value with another advancing memory location playing a role).

Starters are generated the same way for each. No different frame stuffs, all are frame 1, but the XOR was interesting.

Code:

Frame	SEED1		SEED2
0	3D072258	5C841A77	Start 0...
1	F5C0437B	F300431A	
2	957B114A	73342467	End 0 - F301431A
3	86FB0AF5	AC53CDB7	End 1 - 73352467
4	42C20E6C	385453CA	End 2 - AC52CDB7

don't know why, /end

Last edited by Kaphotics; Oct 30th, 2010 at 1:13:03 AM.

==

Coming in to edit months later to clarify on the XOR thingy I said. There's no different XOR after looking at the RNG again. Noobishly I saw it add 10000 for the first and second, and subtract 10000 on the 3rd one. The RNG when advanced and XOR'd does this correctly. No odd XORs here!

Oct 25th, 2010, 1:41:10 PM	#377
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	code discussion, ignore this I made a code The Pal Pad number is not blatantly obvious in the memory (tried full word searching of a 10 byte hex, combinations of them in decimal and hex) and nothing was consistent. Since the pal pad ID isn't squashed together; each 4bit value is separated from each other. It's probably encrypted away anyway, as it's based off of mac address. I wouldn't bother looking further :S I suggest IV MTRNG split into 2 halfwords -> TID + Money White Money is at 0223CC2C TID of Trainer card is at 022696A4 Black Money is at 0223CC0C TID of Trainer card is at 02269684 There is a location for Money displayed on the card, but it freezes DeSmuME if changed at any time. (0226969C white, 0226967C black). Could be the same as the IV check code freeze issue not being a problem on actual DSes. Turning off a code that freezes the value at 022696A4/02269684 (TID) reverts it back to the game's ID (so no side effects to revert if saved). This location is zeroed out until you open the trainer card, so you should activate the code right as you click your trainer card from the start menu (not from when you actually view the card while viewing the badges) Made an IVMTRNG Seed Check code, tested on DeSmuME and on a cart. White Code: 94000130 FFFB0000 DA000000 022151D4 D7000000 022696A4 A226969A 00000000 1226969A 00000000 DA000000 022151D4 D6000000 0226969A D2000000 00000000 Black Code: 94000130 FFFB0000 DA000000 022151B4 D7000000 02269684 A226967A 00000000 1226967A 00000000 DA000000 022151B4 D6000000 0226967A D2000000 00000000 Activate with select once you are viewing your badges, or else the game will freeze or the code will not work. Once activated click your trainer card to view, no need to hold down the activator. It has no negative impacts on your game! Just exit the trainer case and everything is reverted; this code just changes what is displayed, not stored. Money (In Hex): XXXX TID (In Hex): YYYY MTRNG Value: XXXXYYYY It's pretty easy to adapt the code for other seeds, just change the bolded offsets to any other offset to check its value. memory check template White Code: 94000130 FFFB0000 DA000000 _MEMORY_ D7000000 022696A4 A226969A 00000000 1226969A 00000000 DA000000 _MEMORY_ D6000000 0226969A D2000000 00000000 Black Code: 94000130 FFFB0000 DA000000 _MEMORY_ D7000000 02269684 A226967A 00000000 1226967A 00000000 DA000000 _MEMORY_ D6000000 0226967A D2000000 00000000

Oct 25th, 2010, 10:43:35 PM	#382
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	Agreed :) Agreed again :) That code will just check the value at 022151D4 (lol sniped by bond, thx bro) If you went into the template and placed say, 0223EBFC in the code, you'd be able to check the PID of the egg you just received. Use 022160A4 and check it once for a value, and again with 022160A8, you'd check the current RNG Seed for PIDs. What I meant by main RNG is that you could check your initial seed to see if you are hitting the same seed every time to remove erroneous attempts, as well as checking advancement from the initial. ---------- Plugged in some offsets to make moar codes. Included a 1 button activator key if they need to be changed. IV MTRNG Check Code [SELECT] SEED1 Check Code [L] SEED2 Check Code [R] Party 1 PID Check Code [START] codes Replace FFFB0000 etc in the first line with XXXX0000 as follows: Code: FFFE A FFFD B FFFB Select FFF7 Start FFEF Right FFDF Left FFBF Up FF7F Down FCFF L+R FEFF R FDFF L basically yah IV MTRNG Check Code [SELECT] White Code: 94000130 FFFB0000 DA000000 022151D4 D7000000 022696A4 A226969A 00000000 1226969A 00000000 DA000000 022151D4 D6000000 0226969A D2000000 00000000 Black Code: 94000130 FFFB0000 DA000000 022151B4 D7000000 02269684 A226967A 00000000 1226967A 00000000 DA000000 022151B4 D6000000 0226967A D2000000 00000000 SEED1 Check Code [L] White Code: 94000130 FDFF0000 DA000000 022160A4 D7000000 022696A4 A226969A 00000000 1226969A 00000000 DA000000 022160A4 D6000000 0226969A D2000000 00000000 Black Code: 94000130 FDFF0000 DA000000 02216084 D7000000 02269684 A226967A 00000000 1226967A 00000000 DA000000 02216084 D6000000 0226967A D2000000 00000000 SEED2 Check Code [R] White Code: 94000130 FEFF0000 DA000000 022160A8 D7000000 022696A4 A226969A 00000000 1226969A 00000000 DA000000 022160A8 D6000000 0226969A D2000000 00000000 Black Code: 94000130 FEFF0000 DA000000 02216088 D7000000 02269684 A226967A 00000000 1226967A 00000000 DA000000 02216088 D6000000 0226967A D2000000 00000000 Party 1 PID Check Code [START] White Code: 94000130 FFF70000 DA000000 02234834 D7000000 022696A4 A226969A 00000000 1226969A 00000000 DA000000 02234834 D6000000 0226969A D2000000 00000000 Black Code: 94000130 FFF70000 DA000000 02234814 D7000000 02269684 A226967A 00000000 1226967A 00000000 DA000000 02234814 D6000000 0226967A D2000000 00000000 ------ IV's for eggs aren't the same as wild Pokemon >.>

Oct 27th, 2010, 6:21:41 PM	#396
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	Been looking at the memory region for the MTRNG and how it advances, without the CGear ever being on. I started with MTRNG Seed A, zeroed it out, and let it advance. I compared it against a nonmodified. Both were (modified) from the same state around 0.4 seconds before the first change in the value. Tested for different times, so it didn't look like it considered time, or it wasn't updated quick enough to make a difference. Tested on white, inside a cave at victory road, no moving NPCs. Initiating a battle (IVs generated from first line) Code: Viewing: 022151D4 Unaltered / Altered D2A1C834 00000000 -- Start F632D07C B632D07C -- Change 1 looking around for other differences... Viewing: 022158EC Unaltered / Altered 11C08278 11C08278 -- Start BAAA25C7 FAAA25C7 -- Change 1 5CCAEDD2 5CCAEDD2 -- Change 2 now back to the MTRNG... Unaltered / Altered 6A0CFC24 6A0CFC24 -- Change 2 wat, no reliance on the first value? Checked the rest of the memory locations for any difference after the battle screen starts its shifting (2nd change) 022152B8 [BE38DDA3 -> FE38DDA3] 022158E8 [313D1A0C -> 113D1A0C]**right next to the last change 022159D0 [5D619E07 -> 1D619E07] 02215B8C [41543FAC -> 51543FAC] 02215B94 [00000015 -> 00000012] This is the frame counter for the battle, advances 2->270 (622 decimal), with even intervals, with 00000000 not existing -> (270/2-1)16=311 changes which then updates the MTRNG position upon looping. Starts changing once the screen fades to the appearing pokemon from darkness, not when the music starts. 02215E08 [05F23C00 -> 05F93C00] Another advancing thing that I have no clue what it does. 02215F20 [00000000 -> 00B300EC] Rapidly changing back and forth between two values, All 3 values have approximately the same rate. I separated out the last 3 lines as they are not with the block the MTRNG occupies (has 0000000 in between. The last 2 don't look related at all. So in conclusion at a quick glance it looks like MTRNG is completely reseeded in the battle. Not sure if any of the other secondary changes loop back to affect the MTRNG. I'll let it keep going and compare the 022151D4 values against eachother later. 022151D4 could just be a displayed value that is calculated from some other memory.... Does the MTRNG predict IVs for multiple pokemon in a sequence? If so it's most likely a displayed value or reseeded. The MTRNG position advances every 311/60(fps=delay) =~ 5.2s seconds while in battle the moment the battle UI appears. The 270h counter advances at the same rate as the delay location at 023FFC3C does. I guess this information can be useful.

Oct 30th, 2010, 1:20:03 AM	#400
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	When I was doing the set up picture for ya, I didn't see the IV MTRNG location advance at all even throughout the entire process. I'm pretty sure the 022160[A]4(8) location was flying with the rain. Interesting... I'll try this out for the starters to see if they get the predicted IVs. Code: Frame SEED1 SEED2 0 3D072258 5C841A77 Starting Seed 1 F5C0437B F300431A (Nature?) 2 957B114A 73342467 Ending Seed for all 3 starters (F301431A) 3 86FB0AF5 AC53CDB7 Start Belle Battle 4 42C20E6C 385453CA Start Cheren Battle 5 25621B5F DE525286 They left (1) 6 141A423E 8E4E23A4 7 F3A7EF39 A3824B9D 8 96FC8140 31B933B6 9 3BB0DD03 863B8385 10 82496BF2 418042C4 11 BD80B73D 7F700746 12 64A98ED4 1337A036 13 CC476C67 91F60FDC 14 3C288266 E83AD4C1 15 02B4A701 0D93E9CB 16 BBD30B28 7A759954 17 397A6D8B 3F142134 18 D4E0399A E6F2F5C3 They left (2), end (saving now) 19 86A6C285 F437855F after saving 20 CBAD8A3C DCD3BBD7 21 523C446F 80297DDF 22 6596058E 0DFF1D6D 23 6C31CDC9 094383FE 24 C36C6010 375A86C1 The PID is set the second you dismiss the Received Pokemon message, before they start talking to you again. The IV MTRNG changes when you receive your starter. There is no passive advancement for either seed, so hitting a seed and saving to advance is the only way to go for starters. IVs were predicted by the IV MTRNG Seed. The frame advances by 1 when the battle screen starts flashing for each rival. The seed does not change when the MTRNG seed does during the battle. Why does it advance so much after not advancing by much at all? Probably setting your rival's Pokemon, as they were different even with the same seeds, even between save states (thus based off of some other value with another advancing memory location playing a role). Starters are generated the same way for each. No different frame stuffs, all are frame 1, but the XOR was interesting. Code: Frame SEED1 SEED2 0 3D072258 5C841A77 Start 0... 1 F5C0437B F300431A 2 957B114A 73342467 End 0 - F301431A 3 86FB0AF5 AC53CDB7 End 1 - 73352467 4 42C20E6C 385453CA End 2 - AC52CDB7 don't know why, /end Last edited by Kaphotics; Oct 30th, 2010 at 1:13:03 AM. == Coming in to edit months later to clarify on the XOR thingy I said. There's no different XOR after looking at the RNG again. Noobishly I saw it add 10000 for the first and second, and subtract 10000 on the 3rd one. The RNG when advanced and XOR'd does this correctly. No odd XORs here!

Oct 24th, 2010, 11:53:27 PM	#376
Bond697 Dies, died, will die. Join Date: Jun 2010 Posts: 302	nevermind, my mistake.

Oct 25th, 2010, 10:32:00 PM	#379
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	You could probably have 100% accurate tests now with the seed check code changed to check the main RNG. Hmm, it could probably even check the PID of the Pokemon you just caught as well! (goes back to the offsets for party pkm) From my observations of reading the output from researcher, saving increased it by exactly 1, every time. With a different current seed, NPCs behaved differently (fun fact).

Oct 25th, 2010, 11:36:43 PM	#385
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	Yeah Bond now says he's getting confusing results. I'll look into it, it worked fine when I was testing... It works perfectly fine for reading the value on the emulator, and Bond initially said it worked fine on his cart for predicting (he's afk now). I tracked changing memory spots and the card $ and ID are the only ones that are changed relating to activating the code. It works for the party 1 PID check, and the seed checks. I don't see what could be going wrong with the code other than using the wrong version... Did it work at all for predicting IVs? Could just be ingame time/lighting related? Was the game saved very close to the time it was reloaded and checked? Forgot to remove 20 for black for 022151D4->022151B4, FML. That's why mattj's seeds were wrong D:< It's fixed now. I checked to see if the IVs were predicted without the code at 022151B4, and for one test it did. The IVs were the same whether or not I used the code.

Oct 26th, 2010, 8:30:58 AM	#387
ΩDonut sometimes experimentation begins with "let's multi battle strip pokémon" Moderator Join Date: Aug 2006 Posts: 3,691	mattj, I'd like you to try something once you get the revised code inputted. Can you start from October 17, 2010, 18:00:15, and turn on the C-Gear once you enter the game? Not when it gives you the option to turn it on at the start, but in-game with the C-Gear's power button. Record the time you pick yes when it asks you "do you want to turn on the C-Gear?" You should use a timer to countdown when to pick "yes". Give your best guesstimate as to what 2000 delay might be (it takes about 1300-1400 delay to turn on the C-Gear in-game). You know how you add 200 delay* to your HGSS delay in PikaTimer\emloop\whatever to account for the gap between the moment you start the game from the DS menu and the moment it actually starts counting delay? Just use that number. When all that is done, record the RNG result. I have a hunch, but I need to see it tested on a cart. * number pulled completely out of my ass

Oct 26th, 2010, 11:20:22 AM	#389
ΩDonut sometimes experimentation begins with "let's multi battle strip pokémon" Moderator Join Date: Aug 2006 Posts: 3,691	Correct. It's not much different.

Oct 26th, 2010, 9:05:56 PM	#392
Bond697 Dies, died, will die. Join Date: Jun 2010 Posts: 302	i think i've had the cgear on most of the time i've been testing. that explains so much.

Oct 27th, 2010, 12:13:51 PM	#394
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	It's pretty hard for that one since there is the ambient crazy rain (which advances the PID RNG a lot). It appears as this rain won't advance the IV MTRNG. Once you step outside of the house when getting pulled in, you follow a set path where he is generated. Save right before you step outside and you should be able to soft reset em. (if this is what you are asking), or if you want to save him for later just save near the house but don't go near it until you do your soft resets. safe pre-trigger spot You should probably check the mtrng in the crazy rain at different points in time to see if it is advancing, before starting the event.

Oct 29th, 2010, 2:59:54 AM	#398
Kaphotics Remodeling Kitchens Join Date: Apr 2009 Posts: 758 Berkeley, CA	How sure are we that 023FFC3C is the Delay counter? 020A9F60 also advances at a counter like rate, but not the same rate as 023FFC3C. Because the memory location right after 020A9F60 is =[022160A4] (lol we know what this position is!) kinda bored, so I just was looking at how the IVMTRNG was updated and from where in Cheat Engine The 00401000 region (1FD000 long) looks pretty interesting. I can't find it in RAM search (not viewable?). 00429645 is used to write to 022151D4 initially, and 00432383 along with 00429645 is used to write to 022151D4 when the CGear is on. 004329D9, 00432383, 004288ED, 004297AC access 022151D4 when encountering a Pokemon. All of these are also reading 022151D4 when 022151D4 advances during battle. All 4 of these locations seem to "call 004d74d0" with 004d74d0 => sub esp, 14 (83, EC, 14) [subtract 14] along with a few other things.... that's it for now, just scribbling some stuff down.