CRYSTALGATE'S PROFILE

So Leon had almost thrice the HP Carmilla had and he had also far more armor? Yeah, there was no way this could have worked. As far as armor goes though, rather than thinking "the fighter should have x% more armor than the mage" you need to think "the fighter should take x% less damage from physical attacks". This will of course vary depending on enemy unless you're using a divisional defense formula, but an approximation can be made.

Kill a character if the story demands it, but do not use it as a shortcut for anything. Killing a character is not an easy way to get the player to care, rather you have to be good at getting the player to care to make a character death work.

There's always the question of how you deal with the gameplay aspects of that character. I would not suggest a direct replacement. However, I do think you should eventually replace the tactical options that character offers. Tellah was mentioned earlier. Rosa and Rydia has all spells he had. You do not even have to replace the tactical options skill for skill though, just the overall functionality. Eventually other characters should be able to accomplish whatever the deceased characters was able to accomplish.

The elephant wall-paintings are not so much an insider joke as it's a necessity if the game is supposed to take place in Boatmurdered. Anyway, I'll check the game out after rereading Boatmurdered.

I think that if you are aiming for tactics in battles, 3-5 turns would be easiest. With 3-5 turns I mean you aim for 4 turns, but often the player will be able to either finish it in 3 turns only or will need a fifth turn.

Shorter than that and you pretty much kill any skill that doesn't cause direct damage or heal. Longer than that and you most likely end up with duplicate turns, meaning you have turns where the situation hasn't changed in a meaningful way from the previous one.

As has been mentioned, anything can work though if you can figure out a way to make it work. However, most JRPGs have battles where you just spam the same offensive skill and what doesn't work is copying them and expect other results. Also, there is no single idea that will make battles more tactical by itself, all ideas will only have the desired result if implemented in a competent way.

Final Fantasy VII's defense indeed doesn't hurt the game much. It's gives you an opportunity to screw yourself over by equipping a +10 vitality (or spirit) accessory, but makes that scenario extremely unlikely to happen as raising defense and spirit is very unappealing even if you believe doing so would have a meaningful damage reduction. It's a case where a minor design flaw cancels out a much bigger one.

My principle is that if you give the player an option to increase a certain stat at the cost of increasing another, then increasing the former stat should be useful. More precisely, increasing a stat should be useful in a situation where it's implied that it's useful. For example, increasing attack doesn't have to be useful if you build your character as a mage, but increasing defense should be useful when you get hit by physical attacks.

Actually, I think I just list where RPG maker games often mess that up.

RMXP - Dexterity. Dexterity works in a funny way where it tends to be useful only if there is an extreme difference between the attacker's dexterity and the target's agility. Let's say your dexterity is 50, raising it to 75 will be useful if the target's agility is 5 or 500, but not if it's anywhere close to 50. The exception to that rule is if you make a skill dexterity based. Actually, there's an even better exception, that is you redesigning how your game calculated damage and chance to hit, something I'd recommend since RMXP calculates it in idiotic ways.

RMVX - Spirit. Spirit is normally a useful stat in VX, but the SPI-F values of spells in the samples skills are 30 and many people seem to take that as a recommendation. With a SPI-F value 30, spirit is almost guaranteed to become useless.

ACE - Luck. Every point of difference between the attacker and the target increases or decreases your chance to land a status effect with 1/1000. More often that not the the player is only allowed to affect the stat is quantities that are too low to be useful.

All three of them - Evasion. Often done by simple giving the player too low amount. Defense works so that +2 can be useful at the very beginning while being useless at the end. A lot of game creators (I've seen this happen in commercial games as well) handle evasion the same way. Examples I've seen is the player granted the choice between a shield which gives +3 defense and evasion or a gauntlet which gives +3 defense and attack instead. This doesn't work since evasion doesn't scale as attack and defense does, a low evasion is crap regardless of where in the game you are.