So... playing another tactical RPG. Yyyay. Technically, this one's an action RPG with tactical elements, but that's a flaw that I'll just have to see past.
I really like tactical RPGs. While scifi is the setting I'm obsessed with, I'd say that tactical RPGs are my genre of choice. I haven't designed any in a while, because it's arguably the genre Unity is worst at.
Today I suffered the same problem with this tactical RPG that I always have: they keep introducing interesting (or potentially interesting) characters, but I've already settled on a party and don't need them. This is made worse by knowing that even if I ditch an established party member for the new guy, the new guy's personality won't show up at all. They're just bundles of stats.
We want our characters' personalities to matter at the heart of gameplay, and also to push the player to try a wide variety of parties with a lot of different characters. So their personalities need to drive their conflict performance, the way they interact with the world in the core play loop. I'm avoiding the term "battle", because it's actually much easier to do this with a noncombat set of conflicts.
Let's consider the way personality can matter.
The first way is that the core type of the personality can drive interaction. For example, a stoic person and a hyper person would approach a situation from very different angles, and get a different kind of result. This is similar to choosing a technique or move set for a combat class.
Another way is that the mood of a person can drive the interaction. Someone who is stoic or hyper doesn't address the situation differently, but instead feels different things as the situation grows. Those emotions drive their actions, rather than their fundamental personality. This has the disadvantage of being a bit hard for a new player to master, but the advantage of being a whole lot more flexible, tactically speaking. How much of a time delay is required to change moods is a design point: it might be an absolutely immediate response, very similar to how personality alone would work... or it could take longer than the whole battle, requiring careful mood management between conflicts. Or, of course, any length between.
Speaking about time as a matter of importance, we can also talk about not having a single monolithic "personality". Instead, think about it as a series of personality elements. For example, stoic-stoic-romantic, or hyper-lazy-motherly-lazy. Every round, the cursor moves to the next personality element. This creates an ongoing pattern where the overall mood is affected by the current personality element's response to whatever is happening this round. This adds complexity, which means we would necessarily have slower conflicts with fewer participants, but that's not necessarily bad. Differing chain lengths leads to drift, keeping the various participants subtly desynched.
We can also create complexity by layering it. For example, we might layer downward. Below the "stoic-stoic-romantic" might be another set of personality elements. Maybe "romantic-depressed-lazy". In the course of the conflict, you might strip away the surface personality and the one underneath would surface, in which case they might be romantic-stoic-romantic or stoic-stoic-lazy or whatever. This could be considered "damage" if you like, and the way to "win" or "lose" might be to get that lower element stripped away, leaving you unable to properly respond.
We can also layer "upward" or "inward", adding a layer on top of an element, or slotting an element between two elements. These temporary personality elements don't really reflect the fundamental personality of the character, but it is an interesting idea.
At this stage we're still talking as if the conflict will be some kind of winner-take-all battle, with talk of stunning the enemy's personality or forcing them to react a different way. While that may be completely feasible idea, let's also think about what the goal of these conflicts could be. Why would you want to stun the enemy until they are unable to argue against you?
The first thing that springs to mind is that you're deciding something. For example, there's a town meeting and everyone is putting up plans about how things are going to change. The supporters of two conflicting plans argue. The argument is partly about forcing your opponent to see your side and be unable to support their own plan, but the other part of the conflict is convincing the audience that your side is right.
This merges very well with the rest of my design idea, which requires that a steady stream of new characters come in without being simply discarded in favor of your established characters. You're not running a party: you're running a town.
There are no villains, really. Just people who have different visions for the town. On any given day or week, two or three plans will be proposed. The plans always conflict, so you will have to side with one of the teams proposing your favorite plan... or perhaps with the one you don't like, in an attempt to sabotage it.
These will change the nature of the town, surely but steadily. New people will trickle in, and they also find themselves in favor of specific plans, showing up in your weekly town planning meeting. The town itself exists, of course - this isn't happening in a vacuum. If you vote to expand the fisheries, then your town docks will expand, and you'll probably get a few fishermen as immigrants. If you vote to expand the space docks, you'd get more ships landing, and more transient sailors on shore leaves... etc, etc.
However, the town meetings are analogous to the "battle" segments of the game. There would also need to be a lot of between-battle gameplay. And I think this system could support a lot. Ideally you could wander around town as any character, talking to people, chasing sidequests, buying and selling furniture and clothes and stuff, searching for secrets, learning new skills... all the sorts of things you do in RPGs. Plus any multiplayer you may or may not have.
This would help to cement the characters in your mind, as well as allow you to customize them in small ways and feel invested. However, it also allows you to set up the situation for the town hall meetings, because you can get a feeling for what plans are going to go down and who will support what. Then you can attempt to set up certain people with advantageous starting moods, or maybe get other people to take a vacation out of town. Maybe you even start a preemptive argument among someone who is on the fence to get them to join your team in the debate. People's stats will be better the more debates they've had alongside these partners, which is the way you do leveling...
Of course, when it comes to the meeting, you'd have to choose which three of the support characters you'd like to take into the debate, and then juggle them as the conflict goes on and moods slowly shift...
Well, I dunno if it'd be fun, but I think it's an interesting idea.