I feel like I ought to expand a bit on the unusual balance of mechanics that Vagrant Story has, because I think it's somewhat revealing about some of the peculiarities of design relating to jRPGs. So, a wall of text it must be, and a wall of text it shall be:
The thing about Vagrant Story is that it's basically "Fuck You: Suck My Dick: Yasumi Matsuno's Personal Dream RPG Experience", or at least a good half of it before someone told him to actually finish the damn thing. This isn't so strange when you think about it - Square after Final Fantasy VII had, I assume, a preposterous amount of money, so they could authorise all kinds of projects and let their directors have bizarre fever dreams like Xenogears and Chrono Cross. If you think about it, it's not that dissimilar from the sort of managerial laxness that allowed the creation of Fallout and Planescape: Torment at Interplay, except that Square had a lot more money.
Anyway, all Yasumi Matsuno's games have complex and fiddly subsystems that are very difficult to fully comprehend but, if fully grasped, allow the game to be broken apart like a twig. This is not, necessarily, a failure in design - my understanding is that this is the way Matsuno likes it, and because Vagrant Story is a game where Matsuno was, to a large extent, running wild, it encapsulates this philosophy particularly well. So the weapon growth and crafting in Vagrant Story is downright byzantine - the way blade combining works defies logic because in a lot of cases it's utterly arbitrary, and even the growth system, which actually has a system, has a ton of fiddly parts that have a surprising amount of impact. Not only that, if you have no idea what you're doing, it's actually very easy to end up with weapons that are just plain not very good. I know that the first time I played Vagrant Story, probably one in three bosses I couldn't do any proper damage against with normal attacks, and I have very fond memories of trying out various break arts in the Giant Crab's mouth to somehow do more than 1 HP with an attack.
So this is where we get to the issue of crutch mechanics in jRPGs. Vagrant Story is a difficult game, in the sense that it requires a great deal of knowledge to play well, and learning those things by trial and error requires a considerable effort. Basically, the system is so complex that it would be unreasonable to expect players to get the system fully the first time they play; on the other hand, because Vagrant Story, like basically all jRPGs of its era, is also a narrative game, stonewalling the player and forcing them to restart is not an option. So what Matsuno did with Vagrant Story was
not dumb down its immense stat-o-rama, but instead, like I said before, he introduced an alternative system by which you could scrape by even if you didn't understand the crafting system. This, by the way, is one point in which Vagrant Story resembles Dark Souls; the system is so complex that the player is
expected to fail to some degree, and arguably that's part of the intended experience. It's just that Vagrant Story, instead of killing the player restarting them from a check point (which wouldn't really help in Vagrant Story, since the key mechanic is in preparation) lets the player push through bosses using items and chain abilities, but in a way that feels cumbersome, so the player is encouraged to try and do better and create more successful weapons for the next boss.
I think my point is that it is specifically
because jRPGs have crutch mechanics that some of the more interesting jRPGs can have such arcane and complex (and occasionally fun) subsystems: grinding, which is a kind of crutch mechanic in many jRPGs, is there to establish in firm terms that the player
won't be stonewalled, which, essentially, frees them to experiment with whatever obfuscated and only partially informed character development system the game happens to have. On the other hand, success at using the system well manifests in the player
not having to resort to using the crutch mechanic. From this point on, an important point in jRPG design is to balance the two mechanics in such a way that learning to play the game well is rewarding, but without making the punishment for failure
too tedious. I suspect that, overall, the ideal level of difficulty for a jRPG is one that
occasionally pushes the player to the point that they have to use a crutch mechanic, because this keeps the player on their toes while reminding them that they
can make up for failures. In a game that can be around 80 hours long, this is much better than expecting full understanding of the system of the player.
This segues nicely (I think) to what MRY was saying earlier:
One dumb thought I've always wondered about -- something that jRPGs kind of do with how XP rewards and XP level thresholds scale, but in an indirect way -- is whether you could have a system where, upon reaching some checkpoint n, you automatically are raised to level L.n, with L.n being the level at which reaching checkpoint n (typically navigating a dungeon and beating its boss) would be easy to do without intelligent tactics/good luck. This would mean that the reward for completing the dungeon would increase the more efficiently you completed it, and decrease the more you grind. You'd never be leveled down at the threshold, so you could always overshoot the mark and start the next segment with a bit of a boost, but it would avoid the scenario where a player who skips grinding during segment n to n+1 will probably need to grind at least a little between n+1 and n+2.
I'm sure there is some reason why this is a bad idea, or perhaps it's already covered by the XP scaling point above, but it seemed kind of neat to me.
I think the difficulty with this approach - which is not an insurmountable one - is that an XP level is not a sufficient measure of a character's power in a lot of jRPG. Instead, power is a sum of equipment, consumables and long-term character development choices (plus the player's tactical ability). So it might prove difficult to establish a decent "miminum threshold" to upgrade the player to once you pass through a checkpoint, particularly in a way that doesn't downplay the importance of various character development choices and the meticulousness of the player's exploration and so on - basically, the long-term considerations of what the player
does while playing an RPG. I do think this kind of system would work very well in a tactical RPG in which the main course is a linear series of combat encounters, which (one would presume) would have an easily identifiable sweet spot in terms of difficulty. As it is, character development in Valkyria Chronicles is largely a fig leaf to make a wonderful tactical puzzle game
just count as an RPG, and it's still a really good game, so I think this sort of system would suit a game like it very well. For jRPGs that aren't about tactical encounters and more about elaborate character systems, though, I suspect that a character growth mechanic that distracts from the feeling of long-term growth and pay-off from good character development decisions might end up interfering with an important element of what makes the game engaging.