Some interesting RNG info

[VincentASM]

Pikaroom is a pretty cool Japanese site that investigates the behavior of the RNG in various FE titles. Looking around, it gets pretty complicated and talks about RNG mapping and crazy awesome stuff like that.

Anyway, I just noticed it has a couple of sections on this game, and there were some interesting info that I gleamed.

1. Apparently, the game uses 2 RNs to determine Hit rate, like in the GBA series titles.

2. It seems the RN maps are dependent on the DS's internal clock.

3. Complete character growth rates seem to be comprised of 3 factors. The first 2 are the character's inherent base growths and the player class growths, which most people know about. The third one is less well known and seems to be compensation growth rates, which are related to the former 2 factors. The existence of this third factor seems to explain why 0% growth rates may yield stat ups. However I'm not certain whether they exist or not and I don't fully understand how they come about (nor how they work).

Edited December 25, 2008 by VincentASM

[Raven]

Maybe these factors are the reasons that my Marth had 1 resistance by Lv10...

[Dr. Tarrasque]

I've yet to see a compensation growth rate work on something that has 0% but sometimes it seems like a character won't gain a stat-up where he has 90% or 70% just to get a stat-up where he/she usually doesn't grow.

[Rodykitty]

Maybe these factors are the reasons that my Marth had 1 resistance by Lv10...

Marth has 2% Res growth. You really shouldn't be surprised to see him with a Res point.

And I second that I've never seen compensation growth work on a 0%. However, I'm going to boldly assume that the growth rate must be zero percent exact, not a negative growth which most characters have.

...

Also, I already told people before that this game uses 2 RNs (I tested this in the past and was convinced). It's blatantly obvious once an enemy's battle accuracy goes below 50%, and also why dodging was nerfed. People didn't believe me though. I didn't argue further since I didn't have solid proof.

[VincentASM]

It seems compensation growths factor in reclassing, so if you haven't levelled up a character in a different class where their growth is non-0%, their 0% growth will never yield a stat up.

I'll try and test it later, but I need a better understanding of the concept first (since my Japanese knowledge is really crap).

[Rodykitty]

So basically... Growths for reclassed characters are slightly different than we originally thought.

[VincentASM]

Most likely. The tricky thing is that they're apparently additive and thus dynamic. Well, not like people can't factor these additions on.

Also, the additions appear to be very small, like 1%, 3%, etc.

[Rodykitty]

Is this just for 0% growths, or for growths in general?

[VincentASM]

It's for all growths. Hence it shouldn't be too hard to test.

[Rodykitty]

Interesting... Sounds like class swapped characters have slightly better growths than we thought (5% usually makes a difference in averages, so I imagine 3% would as well).

Maybe this is why Jedor's Sheeda kept getting Strength. >_>

Edited December 25, 2008 by FE3 Player

[Celice]

2. It seems the RN maps are dependent on the DS's internal clock.

If memory serves me, this is actually far better than relying on something that movement can affect. While still abusable, it's far more tedious and harder.

[VincentASM]

I thought of one test, but it seems to have failed. I'm not sure if me resetting the game had anything to do with it.

Basically I gave Wrys 0% in all growths and gave Mages 91% in all growths. Assuming compensation growths add 1/10 of the combined growth, it should bump the combined growth to 100.1, i.e 100%. After successfully getting both Skill and Luck to not rise, I saved and levelled up some more. However, there were occassions when Skill and Luck would not rise despite possibly getting a +9% growth boost.

I'm not sure if I've got it all wrong though and I might need to find someone with better Japanese understanding : o

Oh well, time to rethink and maybe reread. I'm thinking the alterations aren't worth bragging about though, but we'll have to wait and see ^^;;;

[Rodykitty]

I'm assuming it's called "compensation" for a reason, since ridiculously high growths wouldn't need to be compensated for.

[VincentASM]

I don't see why they'd differentiate between low and high growths; I assumed it was to compensate for not gaining any stats.

I'm going to try and test a low growth scenario, but I'm trying to think of a good one. Maybe something to do with Jagen...

Also, I thought I started to understand more of what the site was saying, but alas I've hit another obstacle >___<

EDIT

I tested with Wrys again, with 0% growths across the field, both himself and in his class. Prior, I got him to fail to rise in Magic and Speed with his 91% Mage growths. After that, I levelled him up a couple of times, and finally saw +1 Magic.

All I can conclude right now is that there is definitely a possibility of 0% growths to yield stat gains. The correlation to the failed Magic rise prior, does suggest it might compensation growths at work though.

EDIT2

Okay, his Speed went up now o__o

EDIT3

Hmm, after like 50 resets, I got Luck to rise. Uh-oh?

Edited December 25, 2008 by VincentASM

[Rodykitty]

Can you test and see if this works on "negative" growth rates?

[VincentASM]

I tested again with Wrys, but with -5% in all growths for himself and his class. Previously he should have gotten +9% from compensation growths, assuming they exist.

I meant to test compensation growths again, but it was kind of a waste, since Wrys's -10% plus his compensation growths is just <0%. Silly me...

However, after literally over 100 level ups, I didn't observe any stat ups at all (not even from the 0% growth phenomenon). Not sure if it's my bad luck though.

Luckily I had a second test going on at the same time, in case this one failed...

I can confirm that the faceless generics simply use the player class growths for their levels ups and auto-levelling (eg. when you get generics that aren't Level 1). Although most people probably guessed it already.

Here's hoping my third, and more ambitious, test will yield some results D:

Edited December 26, 2008 by VincentASM

[Rodykitty]

I can confirm that the faceless generics simply use the player class growths for their levels ups and auto-levelling (eg. when you get generics that aren't Level 1). Although most people probably guessed it already.

This is nice to know. It would be nice if this was noted on that particular page.