Level 1 Stats

#1venom3384Posted 8/8/2012 1:22:18 PM
Is there any way to calculate a character's level 1 stats when they're not at level 1?

I have some characters at level 99 and I'm trying to decide if it's worth it to go through a leveling/deleveling cycle but I don't really have any point of reference...

Thanks!
---
The Official Venom of C&GN
#2Vir27Posted 8/8/2012 2:49:08 PM(edited)
It'd be a bit tricky if you do not know how what levels you spent in what classes.

If any of your characters auto-joined at high level or otherwise leveled up all in one job class, then, yes, you can use trial and error putting raw stat numbers into a spreadsheet until you find the minimum and maximum raw stats that character could have at his level, and then use those raw stat values to see what range of stats they would have allowed at lvl 1.

The spreadsheet I've been using for years seems to have been lost to the processes of internet geology, so I will make a public google doc for us.

https://docs.google.com/spreadsheet/ccc?key=0AviQt93JXScudEZVU1ZqN2NXZm1pYW5mVVVkaHJ3SlE

So, example: your 99 Balthier has 332 hp naked. As he is male, he has male hp growth and his raw value at lv 1 had to have been between 491520 and 524287. Let's narrow that down.

On the Stats by Lvl sheet, I put 10 and 110 in for C and M respectively (as those are the correct figures for Baltheir's job, as per the second spreadsheet). Feeling hopeful, I put in the max, 524287 under RAW VALUE. I look at the bottom (or make a cell near the top reference the proper cell, to save scrolling).

Putting the max returns 348 hp at lv 99, which is too high. Ok, let's try the minimum, then: 326; too low.

Let's try something in between, then. 500000 returns 332. Great. Let's see what is the lowest value is that still returns 332.

495000 is too low, so we go higher. 499000 is very close. 499500 works. 499300, yes. 499150, yes. 499050, yes. 499020, no. 499030, no. 499040, yes.

So now we know that the lowest raw hp value our boy could have had is between 499030 and 499040. x35 is too low. x38 works. x36 is too low, so we know that the minimum value we are looking for is 499037.

To get the maximum, do the same thing going upwards from 500000.

You see, it's quite simple. But it's only this simple if you can use the calculator, which presently only does a straight stint in one class.

You can narrow down this value even further by checking at each level as you level up, but once you have the 99 figures you can put them in and look at what lv 1 returns to see the minimum and maximum you could have had.
#3venom3384(Topic Creator)Posted 8/8/2012 3:41:02 PM
Dude, perfect. Thanks so much!

Basically, I was trying to correct my Ramza through leveling/deleving since he spent a lot of levels in weird classes to unlock DK and Arithmetics...still haven't decided how I'm going to use him. So I went through a few deleveling cycles as Bard->Mime and Chemist->Ninja but now he's way faster than everyone else...

So, when Reis, Beo, Balthier, etc. joined, they were all at 99, but with speeds under fifteen while my Ramza has 22...

So, the raw stats actually help so I can calculate how many cycles I can do to get their speeds to match and have a proportional number of turns.

Thanks!
---
The Official Venom of C&GN
#4Vir27Posted 8/8/2012 4:07:32 PM
Great, I'm glad it helped. While I was writing it there seemed to me a pretty good chance it would not, lol.
#5RaijiniliPosted 8/10/2012 9:01:31 PM(edited)
Good stuff, Vir, but have you seen the Level Up/Down Simulator?

Not that either is a replacement for the other. Especially since I can't find my copy and I can't find it on the regular place.

Edit: Thought you should know, the corrections (http://www.gamefaqs.com/boards/937312-final-fantasy-tactics-the-war-of-the-lions/43048881/703936317) should be implemented.

Also, I think I've reduced optimal Up/Down to the Knapsack problem. Well, easier than the Knapsack problem, since you don't need an exact fit. I realized that you could figure out the best sequence of cycles to max all stats (to a certain point) by looking at the sum of the logs of the multipliers. Or something.

If no one got that, I might finish writing this spreadsheet and then have to work on figuring out how to present it.
#6Vir27Posted 8/10/2012 11:55:42 PM
[This message was deleted at the request of the original poster]
#7Vir27Posted 8/11/2012 12:15:53 AM(edited)
Raijinili posted...
Good stuff, Vir, but have you seen the Level Up/Down Simulator?


I have not. Also, I have done much much better work for, say, XII:IZJS.

Edit: Thought you should know, the corrections (http://www.gamefaqs.com/boards/937312-final-fantasy-tactics-the-war-of-the-lions/43048881/703936317) should be implemented.


>.> Deja vu.

Edit: wait, so you are confident that these newer stats are right? I see your comments explaining the misunderstanding on some of the other ones, but for Ninja or Sword Saint for example, if you knew this before I am unclear why coldsushi is bringing news. Did you correct this somewhere else and I missed that? But if he is bringing news, and you're suggesting to me it's a valid correction, then how do you know that?

If no one got that, I might finish writing this spreadsheet and then have to work on figuring out how to present it.


I understand the phrase "the sum of the logs of the multipliers" but I do not immediately know hot to apply it, especially with the "or something." Heh. I've never tried level up/down stuff.

Edit some more: had botched my quote tags.
#8RaijiniliPosted 8/11/2012 6:33:28 AM
Edit: wait, so you are confident that these newer stats are right? I see your comments explaining the misunderstanding on some of the other ones, but for Ninja or Sword Saint for example, if you knew this before I am unclear why coldsushi is bringing news. Did you correct this somewhere else and I missed that? But if he is bringing news, and you're suggesting to me it's a valid correction, then how do you know that?

I didn't know about them, but I was double-checking him in FFTPatcher, and also checking the PSX stats in FFTPatcher to see whether they were changes or errors. I thought I might as well have posted what I spent the time to find.

I understand the phrase "the sum of the logs of the multipliers" but I do not immediately know hot to apply it, especially with the "or something." Heh. I've never tried level up/down stuff.

The idea is that you have a start Raw and a goal Raw, and instead of looking at the difference, you look at goalRaw/startRaw (the multiplicative difference?). And you have to realize that when you level up or down, you're really multiplying by a given factor that depends only on your class and level.

So leveling up and then down from a fixed start and end level to reach a goalRaw is really a problem about solving coordinate-wise exponentiation by three factors:
goalRaw/startRaw = (ninjaMult/bardMult)^n*(mimeMult/bardMult)^m*(summonerMult/bardMult)^s.
And multiplication is the exponentiation of addition, so you can take the log of both sides.
ln(goalRaw) - ln(startRaw) = (ln(ninjaMult) - ln(bardMult))n + (ln(mimeMult) - ln(bardMult))n + (ln(summonerMult) - ln(bardMult))n
#9Vir27Posted 8/12/2012 12:49:04 AM(edited)
I don't know how I ended up with that double post... It seems like I must have had to copy my post into a second post and "edit" it. Weird

Right, that makes sense. Are n, m, and s numbers of cycles? How many levels are you doing? I think I've heard 20 and 99, but like I said I've never looked into this.
#10RaijiniliPosted 8/12/2012 10:01:48 AM
Yes, n, m, and s are the number of cycles of each (with Bard).

I'm assuming a fixed level range for each, so ninjaMult (for example) is the multiplicative effect of leveling between 1 and whatever your limit is. Lower levels have a more pronounced effect.