First of all, great work, Djinn.
You guys really should never have let me play this game. Now I'm going to mechanics nerd all over it. We WILL figure this game out in the DL! First, some comments on the discussion here so far...
---
I don't think buffs/equipment to make stats optimised for growth should really be assumed in the DL at all, incidentally. Especially since you aren't optimising them for growth, but instead fixing them equal to enemies, which in some case is making things worse for the PC. Some of the stats posted definitely look a little off to me, and I think it's because of assuming that characters fight enemies with enemy stat modifiers equal to their stats? This will cause the characters to grow further apart in stats than the game allows, which... fits what I'm seeing, the stat spreads you have seem too large in general, except for things which can't grow like a mage's STR.
On the availability note and the decision to give everyone 200 fights, you mentioned that it barely affects humans, but this isn't true for mystics. Sure, this only affects a mystic's CHA (whatever), JP, WP, and HP, but it has a notable effect on those.
For instance, no way should Zozma have 999 HP. Magic users have poor HP growth and his own personal growth is unremarkable. The only reason he could possibly be this high is because you're giving him the same number of battles as anyone else, which... he doesn't join early enough for in either quest he's available in. Ildon and Rei have a similar problem. They're Asellus-only, and you can't get them until you defeat the first four hunters (five in Rei's case), and GreenSage/LionPrincess are certainly not going to be beaten with a party that has fought hardly any battles yet. TimeLord (although it doesn't matter to him really, it does for the averages!) is also affected here, since you always need to do at least 3-4 dungeons before getting him. Rei's HP should be especially atrocious, as 180 joining after Dark Labyrinth is TERRIBLE and she has awful growth. Rouge should be crushing her in HP, and I see where you have Rouge. <.<
---
Anyway, to be more productive, here are some random mechanics stuff which as far as I know the DL doesn't really seem to know about the game, until now:
Deflect: Has a ~25% chance to kick in (assuming STR+QUI+VIT = 200, varies slightly if you see stats higher or lower). In order for Deflect to activate, the character must be using a sword attack that round (which is bad news for Roufas if you allow Deflect for him, since it means he can't dodge while healing). Deflect can also protect allies, but only ones that are "nearby" (mech guide doesn't detail how close) and only if they are unstatused, for some reason. Deflect, in general, can only block singletarget physical attacks, and there are a few attacks it doesn't work on, such as many grappling moves.
Kasumi: Same success rate as Deflect (~25%), but doesn't protect allies (obviously), and only works against melee physicals (again, fails against some anyway, like grappling moves). Again, requires a sword move to be used that turn.
Shields: There are three that are of interest. Note that in order to block an attack with a shield equipped, a character must be doing certain actions. In particular, a few martial arts moves (the important one is Giant Swing) and all magic make shield use impossible. This kinda screws over mages, and completes the triple screwing over of healers who want to evade in the DL - they can't!
a) Buckler: 32% evade against slashing and bludgeoning physicals.
b) Shell Shield: 25% evade against any physicals.
c) Wonderbangle: 98% against piercing physicals. It's very expensive for something so widely wanted (not to mention a little abusive in its spoiling), so probably isn't DL legal to me.
Now, do note that many attacks, in addition to the ones that are just flat-out ITE, have properties where they half or quarter the evade rate from shields (Deflect and Kasumi are NOT reduced in this manner). These attacks are certainly not a majority, but I could see respecting shield evade a bit less because of this, doubly so since they're pretty much a universal boon to all fist/sword/gun users, and possibly some mystics depending on interps.
Regardless, a sword user on the attack is going to evade a bit over half the melee physicals sent his or her way, which is quite impressive (Asellus and Gen are particularly happy).
---
Mystic weapon / martial arts (Last Shot, Giant Swing) instant death:
The success rate of this instant death depends solely on the HP of the enemy. The formula is roughly [Target's lost HP / Target's remaining HP] * 90% + 10%, although do to a bug it's about 1% lower than this (who cares). This means that below half HP, success rate becomes 100%. However, at HP scores above that, the instant death rate falls off fairly dramatically. At 60% HP, it's only barely in turn 1 territory (~69%). At 70% HP, the hit rate is only 48, and against a max HP target it's just below 10%. Note that for all these moves, the instant death check is made BEFORE damage is dealt, so the success rate is based on the pre-damage HP of the enemy. The mystic weapons need to hit in order to make their instant death check, unshockingly. The martial arts moves here are both ITE in all meaningful ways.
It checks normal instant death immunity, for all that other forms of instant death in the game (such as DeadEnd, Implosion, etc.) use a different formula. I don't know what this is yet.
What this means is that mystic weapon and martial arts instant death is actually quite effective... but only against already damaged targets.
---
EDIT: Okay, found the mechanics behind battle rank. Basically, it rises each time you fight a battle. It also initialises to be higher if you have existing game data, but we're assuming that we don't.
Minimum battles fought -> Battle rank
0 -> 1
16 -> 2
32 -> 3
48 -> 4
64 -> 5
96 -> 6
128 -> 7
176 -> 8
224 -> 9
It's not quite that simple. After battle rank 64, what actually happens isn't that the number of battles required for a level increases, but rather the chance that the battle counts towards the total decreases. So there will be some slight randomisation. Also, due to a bug, after about 39 battles at battle rank 9, the game goes back to battle rank 8 for 9 battles, then back to 9 and repeats.
A second complication is some battles randomise battle rank slightly, some areas modify the battle rank of the enemies fought (famous examples are Bio Lab and Yorkland Swamp, which are BR+2) or put a cap on max battle rank. In general, I'd assume these cancel out in the longterm, though - the majority of encounters will be of the expected battle rank.
200 battles might be a slightly low number for endgame since you don't actually reach BR 9, but of course Saga endgame varies hugely by how many optional dungeons you do, and how many encounters you choose to fight in a dungeon. I'd probably feel more comfortable with a BR 9 endgame because mystics are generally seen as having BR 9 monsters (Suzaku for stats, Dullahan for those who allow absorb skills, etc.). Granted, you can get those as early as BR 7 in Bio Lab (+2 BR) so this concern may not hold any weight.
EDIT: For my reference, mostly, since I might as well post these compiled numbers -somewhere-.
Battle rank -> Enemy stat-up value
1 -> 5.5 (16 battles)
2 -> 14.5
3 -> 23.4
4 -> 32.2
5 -> 40.8 (32 battles)
6 -> 49.4
7 -> 57.9 (48 battles)
8 -> 65.5
9 -> 74.0 (26 battles if we go for 250 total)
This progression is roughly approximated by Enemy stat-up value = 8.5 * BR - 2 (this underestimates reality slightly in the middle, but overrates it at the extremes, mostly balances out).
---
EDIT: Ruminating on DSC.
There are five PCs who have talent in 3 of the 4 DSC moves (nobody has talent in all 4):
Annie
Emelia*
Gen
Liza
Red
Emelia's an odd case as she inherently only exhibits talent for 2 of the 4 DSC moves... but in her own quest, she can switch costumes to mimic both Liza and Annie, each of whom have 3, and between them have talent in all 4. (Interestingly, this makes Emelia in her own quest the best learner of DSC in the game, so she miiight have a claim to it, but I'm ignoring this for now since she has no DL claim to fist moves.)
Of the remaining PCs... Annie is a swordsman and is trained as such. Gen is also a swordsman, even though he has a claim to fists via starting with kick - the plot suggests he's one. Red has his Alkaiser abilities, obviously.
Basically, Liza is the only character with talent for 3/4 DSC moves who is a pure fist user, and thus the only one who realistically will spend so much time spamming fist moves as to have a good shot at learning the one DSC ability she lacks talent with. As such she's the only PC I consider with the ability in the DL.
Another, reasonably strong argument is that this is also the "conventional wisdom" of how DSC is treated in the DL. I can't recall serious arguments to take Red or Gen with it (note that Gen is ranked in Middle, in particular). Liza however has been argued to have it even before we knew about innate talent lists in Saga. Granted, this itself probably follows from the former arguments - Liza's sole weapon being fists combined with her 3/4 talent for them (unlike e.g. Fei-on) makes her the most likely to see the ability for many players who do use PCs somewhat similarly to a theoretical stat topic.
Honestly I'm somewhat inclined to extend this benefit to other PCs who lack talent in abilities IF they specialise solely in their weapon. However, they can only gain one such ability this way (any more would take too much time spamming moves to hope to get, and explains why Fei-on doesn't get DSC.) The relevant beneficiaries here are Annie and Asellus, who gain LifeSprinkler; Fei-on, who gains Skytwister (a plot-billed martial artist lacking talent in Skytwister AND the better two DSC fist moves? What the hell, Square); and Lute, who gains Kasumi. This means all the DL swordsmen and martial artists have their best (non-DSC) move, which is nice, and syncs up nicely with Meeple's original stat topic.
---
EDIT: The mech guide provides ways to calculate magic damage. Most stuff in Meeple's topic is correct, but some things definitely need a look at. A big one is Timelord's TimeEclipse, which should hit quite a bit harder than an average PhantasmShot regardless of the stats involved (since the two use the same formula). Speaking of PhantasmShot, there's considerably more variation in the damage of the four moves. I'll have more on this... later.
