Do we know what determines Child Growth Rates?

[Avalanche]

I've heard conflicting theories about child growths in this game. The 2 most common I've heard is that Growth Rates are determined either the same as in Awakening (dads personals + moms personals + kids personals /3) or (varying parents personals + kids personals /2). Can anyone confirm or deny either of these?

Edited February 2, 2016 by Avalanche

[astrophys]

Even I (one of the people who did lots of the experiments for the /2 formula) don't explicitly know for certain…. I don't have an official source that says it, only inference from mathematical predictions over a few trials… nor do I have access to the game code to double check.

Maybe someone else has this information….

[Avalanche]

I gotcha. I'm honestly surprised IS decided to change up the Inheritance so much from Awakening (given what we know about class inheritance now). I've also heard how base stats are determined is different.

[VincentASM]

No promises, but I'm hoping to finally solve the mystery before the North American release.

Before that, I'm working on calculating the child bases.

I just had a genius idea, but hopefully I didn't screw anything up...

EDIT: It worked. The joys of a 99 use Seraph Robe that boosts every stat by 1.

EDIT2: Okay, the 99 uses is a lie. Gotta figure out why that didn't work...

EDIT3: Ohh, of course, it has 99 max uses, but I didn't change its current number of uses >___<

(If child bases have already been figured out, please let me know. But even if it has been solved, I still want to check the formula.)

Edited February 2, 2016 by VincentASM

[astrophys]

Oh, what was your formula, provisionally? We can test it against other data sets to double check….

[VincentASM]

Oh, what was your formula, provisionally? We can test it against other data sets to double check….

I haven't gotten that far yet.

It's a pain to level up during earlygame so the kids don't go over Level 10.

My first hunch was that 4 is the lucky number, but then one of Sophie's stats didn't increment on the mother's 4th stat increase and instead the 5th.

Once I have an easy way to boost the parents' stats, it should be MUCH easier to gather data.

So far I have a 99 use item that boosts every stat by 1, but I need to hack the game again (gonna take another 15-60 minutes to rebuild/relocate the ROM etc.) so I can get a new one with 99/99 uses instead of 1/99.

Edited February 2, 2016 by VincentASM

[Avalanche]

All the luck in the world to you my friend

[VincentASM]

Thanks, I think I'll need it.

I've gathered a bunch of data, but I still need a lot more data to figure everything out.

Right now I can conclude a few things.

In general, the child gains 1 point once the mother or father crosses a certain threshold (possibly the child's bases +4). They also gain 1 point once the mother or father crosses the threshold +4. However the thresholds are slightly fuzzy and further research is required.

So far, it looks like Level 10 children can only gain 2 points no matter how strong their parents are. The data also suggests you can optimise the children simply by training one of the parents.

Edited February 3, 2016 by VincentASM

[Avalanche]

Wow, I can certainly say I didn't expect that. On one hand, I'm kind of against the idea of the Parents stats having little effect on the Children, as it makes some characters feel exclusively like Pair Up Bots. On the other hand, it does seem wonderfully convenient.

[astrophys]

That's great news to hear… hopefully you can eventually get it confirmed….

I guess the next step, assuming you're capable of doing it in the Rom-hacking that you're doing, would be to see if we can change the parental or children growth rates to something absurd (like 100% in each of the stats, or 0% in each of the stats) to isolate the effects of each parent more significantly….

Perhaps give the children 0% growths and then adjust the mother's growths, then the father's growths….

[VincentASM]

Children bases solved! For now : P

Somebody can write down the mathematical formula, but I'll just explain it in words.

First, some data:

Res +1

GR SA OR
9 13 7
9 5 13
9 6 13
9 7 13
9 8 13

SPD +1

GR SA OR
6 10 1
6 4 10
6 5 10
6 6 10
6 7 9

HP +2

GR SA OR
6 14 5
6 10 10
6 11 9
6 12 8
6 13 7

STR +1

GR SA OR
7 11 0
7 10 8

STR +2

GR SA OR
7 15 0
7 10 12
7 11 11
7 12 10
7 13 9

Stat +X is what Gurei gains on top of his base stats. The numbers in the table are Gurei's base stats (for reference), Saizo's stat gains (displayed stat minus class base) and Orochi's stat gains at the exact moment Gurei gains the aforementioned stat.

What happens is that as soon as the mother OR father's stat gain crosses the child's base stat, the game subtracts the child's base stat from both the mother and father's stat gains (if negative, set to 0) and checks if the sum of the mother and father's remaining stat is 4 or above. If it's 4 or above, the child gains +1 and if it's 8 or above, the child gains +2.

In the last example, for when Gurei gains Strength +2.

GR SA OR [SA - GR + OR - GR]
7  15 0  [8 + 0 = 8]
7  10 12 [3 + 5 = 8]
7  11 11 [4 + 4 = 8]
7  12 10 [5 + 3 = 8]
7  13 9  [6 + 2 = 8]

Normally, once the child gains the first stat, you only need to gain 4 more points for the second stat.

However if the other parent hasn't crossed the child's base stat and you train that parent instead, you may need to gain a LOT more than 4 points depending on far away the other parent is.

Here's most of my data: Spreadsheet

At the top are Saizo's stat gains, Orochi's stat gains and Gurei's base stats.

Ignore "robe count", "all stats +1" is how many times the parent has gained 1 in every stat. The other numbers are Gurei's displayed stats.

For reference, Gurei starts with HP +1 and Def +1 before I got him (since I trained Saizo to marry Orochi).

A blue stat is when it first increases.

EDIT: A preliminary formula might be

Child's Lv 10 stats = Class bases + Stat addition + ([Mother's stat gain - Stat addition; 0 if negative] + [Father's stat gain - Stat addition; 0 if negative]) / 4

To simplify the terms, let me doodle on a piece of paper...

EDIT2: Something like this?

Child's Lv 10 stats = Class bases + ( 2x Child's stat addition + Mother's stat gain + Father's stat gain ) / 4 *

* If less than child's stat addition, set to child's stat addition

Need to double check this. Hmm, you can't really express it like that it seems.

Edited February 3, 2016 by VincentASM

[Avalanche]

This is fantastic, great work. I'm still trying to wrap my head around the formulas, but do you think it's possible to get the kids more than +3 to all stats upon recruitment?

[VincentASM]

This is fantastic, great work. I'm still trying to wrap my head around the formulas, but do you think it's possible to get the kids more than +3 to all stats upon recruitment?

Well, you can, but not from their parents it seems.

I think most of child's stat gains come from themselves rather than their parents. Like if the child is over Level 10, they will gain additional stats as if they levelled up when you weren't looking.

If anyone has data for Level 10 kids with high-level parents, feel free to share ^^

Otherwise, I'll start figuring out how to study the child growths later. I like the idea of setting extreme growths.

Right now my plan is to level up Mitama with Paragon and a 99 use 100 exp Bloom Festal or something. Maybe set her caps to 99.

[Avalanche]

Gotcha, cool. Thanks for the awesome work!

[Frozenkappa]

Thank you for all the hard work. I'm sure many people would be delighted to learn this information.

[VincentASM]

OK, I got some data.

MHP STR MAG SKL SPD LCK DEF RES
100 0   0   100 100 0   100 0 Azama's growths
0   100 0   100 0   100 100 0 Setsuna's growths
0   0   100 0   100 100 100 0 Mitama's growths
0   0   0   0   0   0   0   0 Shrine Maiden's growths
50  50  50  50  50  50  50  50 Exorcist's growths

Stat ups per Level:

Mitama
Shrine Maiden
LV HP ST MA SK SP LC DE RE
11 0  0  1  0  1  1  1  0
12 0  1  1  1  0  1  1  0
13 0  1  0  1  1  1  1  0
14 0  1  0  1  1  1  1  0
15 0  1  0  1  0  1  1  0
16 0  1  1  1  1  1  1  0
17 0  1  1  0  0  1  1  0
18 0  0  1  0  1  1  1  0
19 0  0  0  0  1  1  1  0
20 0  0  1  0  0  1  1  0
Exorcist
2  0  1  1  1  1  2  2  1
3  1  1  1  1  1  1  2  1
4  1  1  1  1  1  1  2  0
5  0  1  1  1  1  1  1  1
6  0  1  1  1  1  2  2  1
7  1  1  1  1  1  2  2  0
8  1  1  1  1  1  1  2  0
9  0  1  1  1  1  1  1  1
10 0  1  1  1  1  2  2  1
11 0  1  1  1  1  1  2  0
12 0  1  1  1  1  1  1  0
13 0  1  1  1  1  2  1  0
14 1  1  1  1  1  2  1  1
15 1  1  1  1  1  1  2  0
16 1  1  1  1  1  2  2  0
17 0  1  1  1  1  1  1  1
18 0  1  1  1  1  1  2  0
19 0  1  1  1  1  2  2  0
20 0  1  1  1  1  2  1  1

After tallying, the observed growths are:

Mitama
MHP STR MAG SKL SPD LCK DEF RES
0   60  60  50  60  100 100 0   Shrine Maiden (10)
36  100 100 100 100 147 163 47  Exorcist (19)
-14 50  50  50  50  97  113  -3  Exorcist (19) - 50

Even with the low sample size, there are a few things we can deduce.

1. Azama does jack all.

2. Mitama and Setsuna's growths seem to contribute an equal share.

3. Mitama's effective growths are likely Class growths + [Mitama's growths + Setsuna's growths] / 2

I may grab some more Mitama samples later, but I'll start moving onto Mitama!Kanna next.

[astrophys]

So at least for this example, it seems my hypothesis is correct after all?

We'll have to see how it works for Kanna, as my tests could never apply to her (she's recruited after the chapter, so none of the videos of her recruitment showed her child-sealing).

[Duke of Dozel]

3. Mitama's effective growths are likely Class growths + [Mitama's growths + Setsuna's growths] / 2

So say we take Effie!Lutz. Will the formula be:

Wyvern Rider's 15% + [30% Lutz + 60% Effie] /2 = 105/2 = 52.5% STR growth?

[GoXDS]

So say we take Effie!Lutz. Will the formula be:

Wyvern Rider's 15% + [30% Lutz + 60% Effie] /2 = 105/2 = 52.5% STR growth?

order of operations. it'd be 60%

[VincentASM]

Very preliminary Kanna findings. Was too lazy to hack the ROM again so Kanna and Avatar's growths are unchanged.

Sample size is 19 so far.

STR        15        5            5
LCK        -5    -5            -20
KANNA    30     35     30     40     45     45     25     25
MITAMA*    0    50    50    50     50    100    100    0  
LV    HP    STR    MAG    SKL    SPD    LCK    DEF    RES
11    0    0    1    1    0    1    0    0
12    0    1    1    1    1    0    0    0
13    0    1    1    0    0    1    1    0
14    0    0    0    0    1    0    1    1
15    0    1    0    1    1    1    0    0
16    0    1    1    1    0    1    1    0
17    1    1    0    0    1    0    1    1
18    0    1    0    0    1    1    0    0
19    0    0    0    0    1    1    0    1
20    0    0    0    1    0    0    1    0
Exorcist
2    0    1    1    1    1    1    2    1
3    1    1    1    1    1    1    1    0
4    1    1    0    1    1    1    1    1
5    1    1    1    1    1    1    1    1
6    1    1    1    1    1    1    1    0
7    0    1    1    1    1    1    1    0
8    1    1    1    1    1    2    1    0
9    0    1    1    1    1    1    1    0
10    0    1    0    1    1    2    2    1
11    0    1    0    1    1    1    1    1
12    1    1    1    1    1    1    1    1
13    1    1    1    1    1    1    1    0
14    1    1    1    1    1    2    1    1
15    1    1    1    1    1    1    1    1
16    1    1    1    1    1    1    1    0
17    1    1    1    1    1    1    1    1
18    1    1    1    1    1    1    1    1
19    1    1    1    1    1    1    2    0
20    1    1    0    1    1    1    1    1

MODEL  1    15    48    37    45    48    63    65    13
MODEL  2    15    42    40    45    48    73    63    13
OBSERVED    24    50    29    50    50    66    66    8

Model 1 is { [Mitama's growths from earlier + Kanna's growths + Avatar's Asset + Flaw] / 2 } + Class growths

Model 2 is the same as above but without the Avatar's Asset and Flaw.

It may be too early to conclude, but Model 1 looks like the most likely candidate.

That said, I need a lot more Level Ups. Also, I'll eventually hack the Asset and Flaw to ridiculous numbers and input some more useful class growths to narrow down the exact behaviour.

[VincentASM]

Some good news and bad news.

I leveled up Kanna over 97 times and ended up with this:

OBSERVED    56    26    74    26    -5    98    100    26
MODEL        50    25    75    25    0    100    100    25

From what I've seen, the Avatar's Asset/Flaw has no effect. Str +100 and Luck -100 might as well not exist.

Removing those, Kanna's growths seem consistent with Class growths + { [Mitama's growths + Kanna's growths] / 2 }

Trouble is, there were two anomalies.

1. At one point (1/30), Defence did not go up in a class with 0% Defence growth.

2. Similarly, there was one time Magic did not go up in a class with 25% Magic growth (1/97).

I cannot pin down what could have caused those two occurrences. So it looks like I'll need to change the direction of my research a little...

[lysander]

I'd try it such that the kid's growths + mother's growths would average out to set some class to straight 100s across the board with the currently proposed model. See if you actually observe +1s across the board always, though in all cases another variable could be the actual growth rates not working like we assume they do, similar to SD.

An Oni Savage kid would work best, probably, since their growth rates don't have any 5s so you wouldn't have to figure out how the game is rounding. Outlaw works on the Nohr side. Or villager if you hack that too.

edit: well, guess base class doesn't matter too much in that respect since you could still hit 5s with straight, non-rounded averages but in some cases I think that'd require going over 100 on the parents.

Edited February 5, 2016 by lysander

[VincentASM]

That's what I've kind of been doing; just staring at mostly +1s across the field : P

Instead of setting the total to 100, I've set it to 99 or 101 instead. Hoping to get a +0 for 99 and a +2 for 101. Otherwise if I set it to 100, I may never see any change. So far nothing strange.

I ran into another anomaly earlier though. Mitama failed to proc Defence in a class with 0% growths even though she should have 100% Defence. My only guess so far is maybe kid growths are capped at 99, but we'll see.

Currently transferring another build, which will take another 10 minutes or so... Stupid me, I moved the new file to the wrong folder initially so nothing changed.

EDIT

I'm a double idiot.

I forgot to change Mitama's base growths from 99 to 100. That would explain all of the anomalous stat gains...

Also, I was wondering why Mitama's Shrine Maiden growths weren't changing; I edited the enemy growths, not the player ones. I already edited her player growths all to 0 earlier, so I somehow skipped past them.

That said, I should probably grab a few more results to make sure everything works.

Edited February 5, 2016 by VincentASM

[lysander]

For the child base stats you were only testing for the level 10 bases, right? What if the pattern extends past 8 or greater for recruiting at later story points/higher level kids? I guess you could set growth rates to 0 across the board and use stat boosters to increase parent stats to test it, since in theory that wouldn't let them gain anything variable from the auto-levels.

[VincentASM]

Don't worry, I'll check when I have time.

On my completed Birthright playthrough, I only trained a handful of characters so I have lots of test subjects XD

Anyway, looking a Avatar!Deere's growths, at least two growths are consistent with (Deere + Avatar + Asset + Flaw)/2.

As I suggested earlier, I used a boundary method, setting the kid's effective growths to hopefully be 99 and 101 and watching for a gain other than +1. However it's really tedious since there's only a 1% chance of getting a different stat gain (0 or +2) and in around 95 level ups, only two stats showed both a 0 and +2..

Unfortunately, I kinda screwed up Jakob!Kanna M's growths (I changed Dark Blood instead of White Blood), but checking the results, it just looks like (Kanna M + Jakob)/2.