[FFVII] Chocobo RNG help

[DynamixDJ]

do you know where (if) this is stored in the save ?I Can add that if you do.

It's taken me a few days of learning, wrapping my head around hexadecimal and binary number systems (for the first time in my life I actually understand how 8 bit and 16 bit actually work), and I believe I may have found where the battle count requirement is kept:

0x0D01 - 0x0D11.

These figures were all at 00 before breeding, then several of them increased according to how many chocobos I'd had stored

0x0D0D was at 07 after breeding, reducing this figure enabled breeding of the chocobo kept in stable 4, although from what I can gather the 'can't breed' flag was still enabled (0x0D00). I wasn't sure how to determine the 'battle count required' though. There were 2 other figures set to 7, these were for the other two chocobos that couldn't mate. There was also a decreasing value for the three chocobos, going from 181-179 (or B5 - B3) at 0x0D01, 0x0D08 and 0x0D0C. I couldn't make heads or tails of these, but they ran alongside the '07' figures.

Also, you may want to look into 0x0CAD, 0x0CAE, and 0x0CAF. As far as I can tell, this doesn't mirror 0x04F8 (etc), but instead displays the character that was last in that particular slot when you saved. Possibly.

 

[nfitc1]

Since you're on an emulator, can you track that after each battle and see if it decreases?

As for your RNG quandary, there are likely more checks going on depending on a variety of factors. I'm mostly certain "personality" plays a part in it. The better they "get along" the higher the chick's stats will be (It appears to be a maximum of average PLUS difference [a*b/2 + |a-b|]). The order you pick them in apparently has an effect too. If the first parent you pick has a great personality, it won't matter what the other chocobo's personality is. However, if the first pick is low it will throw as many as two additional RNG rolls to see if they just do get along. If both have a low personality it will roll as many as three times. If you pick the one with the highest personality first it bypasses all of these rolls.
I could be completely wrong. I don't have it in front of me. I'll examine it in more detail later.

 

[DynamixDJ]

This is interesting. I am fairly certain that the stats received (along with sex and breed) have absolutely no difference whatsoever depending on which chocobo is selected (I've just checked, and the three 'variations' that I've tested on so far all had varying personalities). This does help however with the fact that i am having variations on the battle count required depending on which chocobo is selected.

Speaking of which, I tried to 'spot the difference' with two saves, and my head was spinning when I uncovered the 0x0D01 - 0x0D11 range, I'm gonna spend some more time with it today to try and uncover what's going on (the '07' figure appears to be a flag, but I'm not sure)

I did have a few weird things occur. Here's the difference between the two saves that i evaluated:

First I loaded my save. Then I switched Yuffie for Vincent, then I made a fresh save (the party I keep is my way of keeping track of which tests I perform). This is why 0x0CAE had a value of 5 (for Yuffie) instead of 7 (for Vincent) (or possibly vica versa). Then I ran into the stables, fed 62 Karakka Greens, mated chocobo 1 & 2 to fill stable 4 (3 was filled). Then I saved into a new slot.

The following 'peculiar' things happened:

The chocobo in slot one received a penalty in Intelligence (100-50), and a boost in Co-op (0-62)

0x0E41 changed from 00-02 (?)

(This is definitely noteworthy) The values between F06 and F0B changed accordingly:

0x0F06 - 2D - 26 (45 - 38)
0x0F07 - 41 - 45 (65 - 69)
0x0F08 - 4C -4D (76 - 77)
0x0F09 - 45 - 41 (69 - 65)
0x0F0A - 00 - 4C (0 - 76)
0x0F0B - 00 - 45 (0 - 69)

I've just double checked to make sure that these are the 'right way round'. I'd like to spend some more time with these. Perhaps I am looking at them wrong, maybe they are 3 groups of 2 bytes...

The last range of 'peculiar' value changes are:

0x0FE4 - 2D - 24 (45 - 36)
0x0FE5 - 2F - 21 (47 - 33)
0x0FE6 - 2D - 24 (45 - 36)

Also, do you think that the chocobos in stable 5 & 6 act differently in any way due to the fact that they are being stored in a different part of the memory bank?

 

[DynamixDJ]

Ok then Sithlord, I've found the values that affect the battle count, which are all 2 bytes. They do not decrease each time you battle, rather they 'dictate' how many battles are required.

Stable 1 - 0x0D01
Stable 2 - 0x0D08
Stable 3 - 0x0D0A
Stable 4 - 0x0D0C
Stable 5 - 0x0D0E
Stable 6 - 0x0D10


EDIT - sussed it! It's amazing what a good night's sleep can do! 0x0D01 is 16 bit, and is the figure your battle count needs to be before mating to continue. So, the value of 0x0D01 minus the value of 0x0BBC (current battle count) is the battle count required before mating can continue.

i hope this is enough to go on, as having the ability to see the count required will really help my my rng testing. Also, is there a way to view the chocobo 'type' once it's in the stable? I can help you find that aswell, this little venture has been fun! Lastly,

CONFIRMED - 0x0CAD - 0x0CAF is NOT a mirror of 0x04F8 - 0x04FA. It is indeed, the party member that was in that slot the last time you saved (or simply just the last person in that slot? more testing required).

 

[nfitc1]

This is interesting. I am fairly certain that the stats received (along with sex and breed) have absolutely no difference whatsoever depending on which chocobo is selected (I've just checked, and the three 'variations' that I've tested on so far all had varying personalities). This does help however with the fact that i am having variations on the battle count required depending on which chocobo is selected.

Perhaps not directly, but the RNG should be moving differently depending on what the personality of the first parent is.

(This is definitely noteworthy) The values between F06 and F0B changed accordingly:
0x0F06 - 2D - 26 (45 - 38)
0x0F07 - 41 - 45 (65 - 69)
0x0F08 - 4C -4D (76 - 77)
0x0F09 - 45 - 41 (69 - 65)
0x0F0A - 00 - 4C (0 - 76)
0x0F0B - 00 - 45 (0 - 69)

These values are peppered all over the chocobo scripts, both in breeding and moving to pens. I don't know what these mean.

Also, do you think that the chocobos in stable 5 & 6 act differently in any way due to the fact that they are being stored in a different part of the memory bank?

I don't think they're treated any differently. I think they just got shoehorned in late in development because breeding certain chocobos with just four pens is a little tricky (maybe).

EDIT - sussed it! It's amazing what a good night's sleep can do! 0x0D01 is 16 bit, and is the figure your kill count needs to be before mating to continue. So, the value of 0x0D01 minus the value of 0x0BBC (current kill count) is the kill count required before mating can continue.
...
CONFIRMED - 0x0CAD - 0x0CAF is NOT a mirror of 0x04F8 - 0x04FA. It is indeed, the party member that was in that slot the last time you saved (or simply just the last person in that slot? more testing required).

Kill count? That's new. If this is the case, couldn't you just have a character with a low kill count make a chocobo mate, switch to a character with a high kill count and immediately mate again?

i hope this is enough to go on, as having the ability to see the count required will really help my my rng testing. Also, is there a way to view the chocobo 'type' once it's in the stable? I can help you find that aswell, this little venture has been fun! Lastly,

As far as I know, no. The only difference is each type of chocobo has a range of stats. Goods & Greats share a certain flag somewhere and Wonderful has a different flag. The others don't have any flag. It's only used to determine what color chocobo results from mating. Certain nuts even have different effects on stats, but that's not well documented.

 

[DynamixDJ]

Kill count? That's new. If this is the case, couldn't you just have a character with a low kill count make a chocobo mate, switch to a character with a high kill count and immediately mate again?

No. 0x0BBC is the total number of battles fought, the same figure that the sleeping man states. I messed up my wording (now edited).

As far as I know, no. The only difference is each type of chocobo has a range of stats. Goods & Greats share a certain flag somewhere and Wonderful has a different flag. The others don't have any flag. It's only used to determine what color chocobo results from mating. Certain nuts even have different effects on stats, but that's not well documented.

If this is true, then when it comes to doing my 2nd generation tests (later on today) there shouldn't be any variation on the outcome of the offspring based on which 'type' of parent is used, which will be nice  There has to be a flag somewhere though that states which type a yellow 'stabled' chocobo is, I'll try to find it.

Also, the reason I am under the impression that the 'coloured' offspring is assigned a type, is because of Terence's Chocobo breakdown https://www.ff7catalog.com/threads/1828/ but after having just re-read it, I realise now that it is slightly misleading
Code: [Select]

Otherwise, if both of the parents are Great or Good Chocobos:        If the total number of races won by the parents is 4 or more, then          the baby is automatically either a Blue (50%) or Green (50%)          Chocobo        If it's less:                   69/256 chance: Blue Chocobo                   69/256 chance: Green Chocobo                  118/256 chance: Yellow Chocobo        The Rating of the newborn Chocobo will be Great if both parents have          the same Rating, and will have an equal chance of Great or Good if          the parents were of different Ratings

it doesn't actually state that the rating only applies to the yellow chocobo born. I'm glad that's been clarified!

 

[nfitc1]

From my own testing, chocobo offspring have no "rating". They are colored and inherit type flags from their parents. If you mate two goods, you'll get a good. Two greats make a great. A good and a great might be a good or a great (I still think they share the same flag). A wonderful mated with anything else other than a wonderful will lose the wonderful flag.

I think of it like this (but it's probably not like this:

Code: [Select]

Chocobo type:0x00 (Yellow, nothing special)0x01 (Yellow Good/Great)0x02 (Yellow Wonderful)0x04 (Blue)0x08 (Green)0x0C (Black)0x0E (Gold)

I have no confirmation of this so I can't point to anything that says this is correct.

 

[DynamixDJ]

From my own testing, chocobo offspring have no "rating". They are colored and inherit type flags from their parents. If you mate two goods, you'll get a good. Two greats make a great. A good and a great might be a good or a great (I still think they share the same flag). A wonderful mated with anything else other than a wonderful will lose the wonderful flag.

FOUND IT!!! It took me a while, but (at last) I can safely say after having ran a few tests that:

0x0E3E - Type of chocobo in stable 1 (1=Wonderful, 8=Terrible)
0x0E3F - Type of chocobo in stable 2
0x0E40 - Type of chocobo in stable 3
0x0E41 - Type of chocobo in stable 4
0x0E42 - Type of chocobo in stable 5
0x0E43 - Type of chocobo in stable 6

I am fairly certain that the Great and Good do *not* share the same flag, as if they did then there would not be any variation with my RNG testing when varying 2xgreats, 2xGoods and 1xGreat with 1xgood.

Also, Personality has no affect whatsoever on the sex, colour or stats that the newborn receives.

So Sithlord, there's two ways in which the chocobo section in BC can be improved (which btw I'd be very grateful if you could include 'type' of chocobo in the stables, along with battle count required, it'll help me on my RNG quest, although now I know where the values are myself....)

Is there anything else that I could look for? I'm on a roll.

 

[nfitc1]

Sorry I've misdirected you at almost every step. I either know what I'm talking about or I am completely wrong.

So do offspring have a chance of being a better quality chocobo than either parent? What quality are the colored ones? More importantly, what does quality do other than cap stats and determine what color comes out?

 

[DynamixDJ]

It's all good, if it wasn't for you I wouldn't have a clue how to read MR 

By 'quality' I'm going to assume you mean 'type', according to everything I've read, type means very little after the 1st generation. If you are able to breed a Wonderful Gold Chocobo, then you 'qualify' for the Gold chocobo breeding circumstances when breeding the Wonderful Gold with a Black (if the Gold was a 'Great' Gold, then you could still get Gold, as when you fail to qualify for the 'special breeding circumstances', you will have 50% of father/mother's colour).

Breeding a Great Gold with a Great Yellow (with 4 races) would result in a Green/Blue. I'm not 100% on all this, this is just from what I've read, but now I know where on the hex editor I can monitor which type a chocobo is, so I'll be able to test in due course

In terms of my experiments, I beleive I may be in for a lot of testing. I'm mapping out every outcome of each RNG's position (all 256 of them), for every variation of chocobo. I'm currently on 1Good, 1Great 0 wins. Now that I know that the chocobo's type varies with Green & Blue, I could be in for a ridiculous amount of testing. I.e.

Great Green M with Great Blue F
Great Green F with Great Blue M
Great Green M with Good Blue F
Great Green F with Good Blue M
Good Green M with Great Blue F
Good Green F with Great Blue M
Good Green M with Good Blue F
Good Green F with Good Blue M

And then I have to throw Wonderful into that equation. Ugh! I'm thinking, is there a program available that would map out the result of each RNG position? It may be less time consuming that way, even if I do have to decipher the hex positions and values. -edit- I've just worked out that there are 18 2nd generation breeding variations, which is 4608 RNG outcomes!!!

i won't know until I get there, but I may get lucky and find that the RNG sequence for stat distribution (as an example) would be the same as the previous generation, and only the colour varies. It's possible..... (fingers crossed)

 

[DynamixDJ]

As for your RNG quandary, there are likely more checks going on depending on a variety of factors. I'm mostly certain "personality" plays a part in it. The better they "get along" the higher the chick's stats will be (It appears to be a maximum of average PLUS difference [a*b/2 + |a-b|]). The order you pick them in apparently has an effect too. If the first parent you pick has a great personality, it won't matter what the other chocobo's personality is. However, if the first pick is low it will throw as many as two additional RNG rolls to see if they just do get along. If both have a low personality it will roll as many as three times. If you pick the one with the highest personality first it bypasses all of these rolls.
I could be completely wrong. I don't have it in front of me. I'll examine it in more detail later.

I stand corrected, you are right. I ran a few basic tests before, but I'd overlooked something. Going back to testing personality, it does affect the RNG by 'offsetting' the value by as much as two, however the 'sequence' remains the same (thankfully). However, the parent that I selected first made no impact whatsoever on the results.

-edit- I'm wrong again, 3 different RNG sequences are used depending on whether or not the chocobos are both personality 0, either of them are personality 0, or neither of them are. I can't be 100% sure at this stage though, but that's how it appears.

This is a huge spanner in the works

 

[nfitc1]

Hey, have you found initial stats yet? It just occurred to me that the stat assignment is probably in the world map scripts. There's no editor or viewer for those scripts. MOST of it has been reversed, but I don't know how to determine what chocobos are assigned.

I say this because I know the same battles always yield the same class chocobo (eg. you won't find greats in icicle land and wonderfuls only appear with one or two "Jumping" enemies). I'm pretty sure the whole chocobo battle is a world script. You can't get one without the chocobo lure materia; The chocobo battle actors all share the same AI scripts and they don't assign type or anything; The "send to ranch" menu has to be a script. I'll take a look and see if anything pops out.

 

[DynamixDJ]

Judging by what I've read in Terence's Chocobo breakdown, the fact that I can't find the values (although I haven't looked too deeply), and my RNG results, I'd say that a penned Chocobo has no stats assigned to it other than 'type', but I could be wrong.....

In fact, if the stats were determined at the point at which a chocobo is caught (yes I am aware that MDash and Stamina are initially determined (from 8 varying possibilities), then every stat (including MDash and Stamina) is then decided/adjusted accordingly), then there would be some inconsistencies with my RNG results. I've just ran a few tests, I put one Great chocobo in my pen via BC, then I caught a 3 Great chocobos, one alongside two Spirals, one alongside 1 Spiral, and one alongside 2 Valrons. The stats that are given was the same for all four, when moving @RNG-002.

I'm using a great program at the moment called Cheat Engine, I have been able to assign the address for each of the chocobos stats for each of the 6 stables, so I can see them live. I can also edit the RNG value live, I'm currenty assigning a load of desired values, such as 'Love points', hell I can even put a chocobo into the pen without even having to exit the game! It's brilliant!!!!

 

[nfitc1]

OK. If that's not the way it works I'll have to keep looking.

I only suggested the world scripts because I know that's how type gets defined before penning it. I happened across the code that assigns battle formations on the WM and noticed that there's a chunk for chocobo formations including type of chocobo. This comes from the enc_w.bin file, which is documented, but I only recently noticed it.

Cheat Engine is great. I use it along side lots of programs I try to modify. It's great at most stand-alone games (even newer gen ones like Dead Space and FTL). Not so hot on flash games though. Using a combination of that and a debugger called IDA I can reverse engineer/modify the game to do nearly anything. I made a No-CD patch with its help that only changes eight bytes in the exe.

 

[DLPB_]

Finally, CE 6.5 is out!  Will fix a load of issues I reported

 

[nfitc1]

Wow! I knew there are tons more features to CE than I've been using, but I had no idea there were still so many features it could use! If I did anything other than increase money/HP/MP/Ammo/etc with games then I'd care more.

Still, good to know it's getting even more powerful.

Edit:
They need to proof-read their release notes a bit more:

Rightlicking on "your system supports dbvm" will let you manually load DBVM for each cpu. This is usefull if for some reason your system crashes when it's done too quickly

 

[Bain86]

About Chocobo Breeding.. as to, how to get a blue or green at your first try. I have found that the order of what gender you pick first to breed, "Male" then "Female" = blue.  "Female" then "Male" = Green. I would assume the gender must have to change by what spot the chocobo ends up in the stables but i dont know forsure.

 

[DynamixDJ]

^^Replying to a post made over a yr ago lol. It's amazing how much I've learned since I started this topic (this is where it all began for me, right here).

It's safe to say, that with the help of NFITC1 I'm completely sure of the breeding mechanics, and there is no doubt in my mind of how it works now. The first Chocobo you select does matter, but only in certain circumstances. It is not as simple as saying that picking "male" then "female" will result in a blue; more likely that the first Chocobo you select will affect the RNG roll, and will give you a different colour accordingly, but that colour is not set in stone. The RNG marker position at the time of breeding ultimately dictates the outcome, of which there are 256.

Note that the order of the Chocobos that you pick will only affect the colour of the Choc, the stats of the newborn remain consistent regardless of which one you pick 1st (again, the RNG marker position dictates all)

 

[DLPB_]

Is there no way to just add a few calls to the RNG opcode in field using makou?  Could make it more random using certain tricks like feeding the timer into it?  Havent looked at how it works.

 

[DynamixDJ]

Yes, I'm sure there are too. I know that when a random number is generated, the RNG marker advances by one. I think that there are numerous random numbers pulled with certain character animations; I planned to look at the Chocobo Square or Wonder Square to fully work out the tricks they are using, then try to apply that to Esther.

Let me know if/when you want me to look into it for you.