Using Top-Bottom or Bottom-up approach to design data

I tend to use the theme approach to build the framework for abilities and then use a "baseline" as the starting point for each model/character/etc. For example, a regular soldier has these standard abilities or stats; the bird-people are fast so they will have a higher movement stat, the rock-people have a better armor stat, etc.

I rarely use the permutation approach as I think it adds too much of a "calculated" feel to the game. However, having said this, I do agree that a mathematical analysis is good for catching oversights or statistical design problems.

I agree with your pro's and cons, with maybe the addition of some consideration about how the players see all this in the game. Designing the system is one thing, but how the players interpret it may be entirely different. One point of using the mathematical design could be to make a more consistent, predictable world for the players in a way that keeps them drawn into the game without distracting them with obviously "gamey" mechanics that may result if you paint yourself into a corner.

When I reached the end of your post I was wondering if you had considered the hybrid approach, but it appears that you have.

I have been wrestling with this for my own design, with a tug of war going on in my head about which of the 2 options I should put more focus on.

In the end(or maybe for now), I have chosen to build my foundation primarily mathematically, instead of thematically.

The reason for this is that the foundation of my systems should be extremely stable. I should be able to rely on it to work roughly the same way, no matter which theme I throw at it. This allows me more design space when creating themes since I know the core of the game can support it without tearing itself apart.

For example:

7 Domains(think super Color Pie), with 7 classes in each Domain(each tied to one of the 7 primary attributes). For instance, 3 of the domains are Physical, Mental, and Spiritual and there is 1 Power Attribute class(Warrior, Arcanist, and Warlock respectively) in each of those domains. Each of those Power classes is roughly equivalent stat-wise, but where they differ comes in the theme which I have great room to manipulate since I have a stable, balanced Core.

Additionally, the 7 stats(3 offense, 3 defense, 1 energy) are all balanced in relation to each other, with no gaps.

Other systems are also mathematically balanced such as the weapon/armor system, which acts as a secondary class structure. With the core of the game mathematically balanced, it leaves more room for me to play around with the themes. Pull off the same results with different play styles and basically spend most of my efforts balancing at the point of creation instead of all throughout my systems.

The theme will play itself out in the races, elements, nationalities, etc of the world I am creating.

Elements are actually what was tying me up for months, because they are so thematically driven. Elements such as fire, ice, spirit, chaos, plant, animal, explosives, etc, etc. The possibilities are endless. I could never tie it down and be satisfied that I had covered all the bases(couldn't keep track of all the branch abilities) so I had to flip my system on it's head.

Now I don't have to keep track of branch abilities, as each element stands on it's own. A pyromancer and volcanist may both use fire, but in the end they can go about their business in completely different ways. Both may be tied to the same core class and weapon, or they could use a completely different base. Either choice is only relevant to the theme, and not my core balance.

By primary and secondary objectives, you mean similar to the ways Blue can win in MTG? Mill their deck, control all their permanents, reduce their life to 0, etc?

Basically the color pie from MTG. This faction gets these, and not these. Or it gets the others, but it costs more resources or some other sacrifice. Kind of like creating asymmetry within your symmetrical system.

I think symmetry is key at least in some areas for a few reasons. Balance is the obvious reason. Another is that humans are drawn to symmetry, since most of the world(especially biological) is symmetrical in some way or another. Another is that symmetry can simplify some of the more complicated mechanics that you don't necessarily want players to either focus on, or spend too much brain power or game time thinking about. For instance, all characters use life points. If you mess with that symmetry be prepared for players to be asking all kinds of questions. If you planned for that, that can be good asymmetry, but if not then you may be over complicating your game.

I believe asymmetry is best used on the aspects of your game that you want to be most appealing and take up most of your player's time. The Melee/Healer/Ranged trinity is a classic example of asymmetric design at work. 3 unique concepts that can be hard to balance but are designed to draw and/or keep the player's interest. There is a lot of symmetry underlying that Trinity of course. Skill points, Life Points, Damage Types, etc. Mana, unless the designers want to introduce asymmetry by giving the classes different types of energy.

In order to keep the complexity of the game in check, you may make one system symmetrical just so you can make another asymmetrical, if that is where you want to draw the attention.

For your CCG, you might google for a list of all the MtG abilities, keywords, etc. Compile all of those mechanics into a spreadsheet and think about mechanics that the MtG has missed or simply not implemented yet. Even with all that MtG has created over the years, there are tons of abilities that you could add(or even subtract) to a CCG to go above and beyond MtG, especially if you have other unique systems to play off. There are also plenty of other CCG's to draw examples and inspiration from, though I don't think any are as extensive as MtG at this point.

a hybrid if you use it right

the bottom up approach MUST be used to design your core cards, the basic of the basic, the cards that express the mechanics of your game

you can go top down when you get the core cards right, then you experiment with themes and effects since you have a yardstick to go by already