In Secret of Mana, two of your three characters have magic, up to a maximum of 99 magic points at a time. Your offensive caster has an MP absorb spell, but aside from that, the only way to restore your magic points mid-dungeon is to use a Fairy Walnut, which restores 50 MP, and you can only hold a maximum of four.
By limiting the selection of Mana restoring items to one large chunk, the game has rendered it almost pointless to increase your maximum mana above 50. Since your MP caps at 99, you can never use two in a row without wasting some, and given the real time nature of the game, using an item doesn't cost a turn, or similarly penalize you time wise. Your offensive spell caster has an MP absorb spell, so you only use restorative items on your defensive caster, in which every point of MP gained up until 50 is actually worth 5 MP (the original plus four restores), but past 50, every point of MP gained is worth only a single point, to be used and only restored once you sleep at an Inn.
You have no control over your advancement, so this isn't a balance issue, but you can tell from looking at the game that the designers didn't think much about that. Health restoring items come in 100, 250, and “all” versions (out of a max of 999), so why not magic?
“Video Game Design” as a valid field of inquiry is a fairly new phenomenon, and I think design hiccups like the above one from Secret of Mana suggest that it's taken a while to connect the people who were pioneering this new medium with the people who had been doing the same thing on paper for at least a decade before.
Early RPGs (which I count as everything leading up to the release of the PS1), particularly Japanese ones, provide a fascinating window into a large group of people trying to reinvent the wheel in terms of roleplaying games.
For instance, take the idea of critical hits. (Wikipedia article here) The idea of critical hits, as best as I can tell, arose as a way of somewhat mitigating the horrible inaccuracy of the HP system. There's always a chance you'll stab somebody in the heart, causing much more damage than if you stabbed them in the thigh.
According to the Wikipedia article, 1975 was the birth of critical hits in western civilization, which was drastically before the advent of video roleplaying games. Supposedly, the man behind Dragon Warrior, Yuji Horii, cited Ultima among his influences for the gameplay of Dragon Warrior, but even crafting that sentence indicates how much of the early Japanese RPG experience was informed by Western Culture. The Dragonlord from Enix's first Dragon Warrior is most definitely the kind St. George would have been tasked with slaying.
And yet, take the thief class from Final Fantasy. If I recall correctly, he had an increased chance to critically hit, which makes perfect sense to our current gameplaying sensibilities, but actually has no realistic basis. Why should I be more likely to hit a vital organ because I steal things? I understand that there's an association between stealth and assassination, but I imagine that the original Japanese word for the “Thief” in Final Fantasy provided a smoother segue to the upgraded “Ninja” class.
The original “Rogue” (Wikipedia article here) had no such association between thievery and likelihood of striking an internal organ, and yet it's become such a huge part of our gaming culture that nobody bats an eye at the idea that a primary statistic designed to represent your ability to dodge also increases your chance of striking an internal organ, or that there's an entire class in almost every single MMO created based on the idea of the combination of both being sneaky, and critically hitting someone. I certainly don't object, but I think it's occasionally edifying to trace the history of certain gaming tropes, and learn how potentially artificial these associations are.
Well, at least for you guys, I roll combat with my rogue.
Join me for the entire month of June, as I do a series on the tropes of early console and computer RPGs.