Credits: Nitrodon (dynamic growth reduction)
Most people would be happy to use the following formula to calculate the total, effective growth rates for a character:
Total growth rates = character’s personal growth rates + their class growth rates
However, this is not the complete story.
There have been many reports of Level Up stat increases not conforming to the rule. For instance, calculated growth rates of, say, 90% have been reported to yield +2 in a stat. Even more odd, people have seen growth rates of, apparently, 0% yielding +1 to a stat. Naturally these occurrences are impossible if everything is as we believe it is. So where have things gone wrong?
The first place most people would look at is perhaps the growth rate values themselves. Maybe the character and/or class growth rates are wrong and are actually higher than what we believe?
Yet, that is not true. As one of the first people to hack the game and obtain the growth rates for this game, I can assure you that the character and class growth rates listed on this site are all correct (well, with 99% certainty at least). This is because the growth rates are very easy to find and very difficult to get wrong.
Dynamic Growth Rates
If the growth rates are correct, the next place to look is the formula itself. Assuming the parameters aren’t multiplied by any weird numbers (eg. 1.1 × Personal growths + 1 × Class growths), then there must be some extra term(s) at the end.
This is where dynamic growth rates come in (if you can think of a better name, please tell me).
Total growth rates = character’s personal growth rates + their class growth rates + dynamic growth rates
Dynamic growth rates are added whenever a Level Up occurs.
If a stat fails to increase during a Level Up, 10% of the stat’s total growth rate [not counting existing dynamic growth rates], at the time of failure, is added to the stat’s dynamic growth rate. So if the total growth rate was 30%, the dynamic growth rate added would be 3%. If any decimals are present, the game rounds them down for Level Ups, but keeps a record of them for future dynamic growth rate addition/subtraction.
Conversely, if a stat successfully increases during a Level Up, 10% of (100 – the stat’s total growth rate) [not counting existing dynamic growth rates], at the time of success, is deducted from the stat’s dynamic growth rate. However if the dynamic growth rates are negative, the game will treat them as 0 during Level Ups, but will remember their true value for future dynamic growth rate addition/subtraction.
As one can see, the existence of dynamic growth rates leads to character’s growth rates subtly changing all the time, hence the term “dynamic”.
These extra growths can account for why 0% growth rates yield stat ups. Dynamic growth rates factor in reclassing- so if a character Levelled Up in a class with a total non-zero growth rate in a stat (say, a Res growth of 10%), it would be possible for them to raise that stat in a class where it is 0% (in this case, the total, effective growth rate would be 1%, but it’s still more than 0%).
For +2 stat ups, it’s slightly trickier. Since you’d usually need to have many Level Ups with failed stat increases in order to gain a +2 in the end.
To be honest, I don’t really know if I made any sense so maybe these examples will help.
Ex. 1: A simple Level Up
Let’s use Marth for this experiment example. Everyone uses him, he can’t reclass. Perfect.
Marth’s total growth rates are:
Marth Levels Up and his HP, Str and Lck increases.
In that case, these will be his dynamic growth rates:
Which changes his effective growth rates to:
For HP, Str and Lck, his dynamic growth rates are negative, so the game counts them as 0 instead for the time being. His real Res growth is 2.2 (2 + 0.2), but the game rounds down decimals for Level Ups. Skl, Spd and Def are the interesting stats because they have received notable dynamic growth additions, eg. 40 + 4 for Skl.
The result of this Level Up means Marth has a slightly higher chance of increasing Skl, Spd and Def in his next Level Up, thanks to not increasing them in this Level Up.
Ex. 2: How to get “impossible” stat ups by reclassing
I’m not feeling very imaginative right now, so I’ll use Frey for this example. Let’s reclass Frey to a Curate. His total growth rates as a Curate are:
Frey Levels Up and his HP and Lck increase.
His dynamic growth rates will therefore become:
Now let’s reclass him to a Cavalier because that was a terrible Level Up…
His total growth rates [ignoring existing dynamic growth rates] as a Cavalier are:
If you factor in his dynamic growth rates, that turns his total, effective growth rates into:
As you can probably see, if dynamic growth rates didn’t exist, it would be impossible for Frey to increase Mag or Res as a Cavalier. However, if you do the right things, you can make it possible.
Well, I’ve possibly said a lot of radical things up to now, but… er… is it all true?
That’s actually a hard question to answer, since the dynamic growth rates I’ve described are entirely empirical. I haven’t looked directly into the game’s code to find out, nor have I looked it all up in a guidebook. The developers of the game haven’t even explained how growth rates work in this game, let alone mention dynamic growth rates. So it’s up to us players to figure out the whole story.
What I have done is hack the actual growth rate values and test how Level Ups are affected (which is admittedly something the average player probably cannot do). I noted down which stats rised or not during each Level Up and then chose a particular Level Up to repeat dozens of times, to see all the stat increasing possibilites. This allows a determination of the total growth rate at that point, which in turn tells you about the dynamic growth rate value. I’ve tried to be very thorough and consistent with my tests, but it’s still possible that I’ve made mistakes or misassumptions.
In any case, a handful (literally) of my results can be found here. There was more data, but it was lost following a rollback (and I’m too lazy to re-format it).
Otherwise, before I even knew dynamic growth rates existed, Pikaroom (a Japanese fansite) performed their own tests, after noticing the total growth rates were sometimes different to what they believed, whilst generating Random Number maps by manipulating the DS’s internal clock. Their results formed the basis of my research, although my description of dynamic growths is slightly different to theirs (mainly the reduction part).
Just like the case with Path of Radiance‘s Fixed Mode, I don’t think anybody is seriously expected to keep track of all the dynamic growth rate changes all the time. In fact, I don’t even know why they exist. I’m assuming it was some master plan devised by Intelligent Systems to screw us up or something.
Anyway, there are still a few more things to look into (such as growth rates near 100%). However I suppose the current description of dynamic growth rates will suffice. For now, I think the most important thing is to let people know that dynamic growth rates, at the very least, exist.
That way, I won’t get people complaining to me that the growth rates are wrong : )