I've come up (yet again) across this issue where I need to define a set of finite 'resources' in an non-finite environment that has expansion plans.

Specifically:

In a tactical 2D turn-based combat skirmish game, I need to determine damage 'types' (piercing, slashing, fire, etc.)

Every 'unit' I will add later down the road will be limited by these damage types.

I'm trying to find a good fit by keeping things simple, but I can't help thinking I'm inevitably closing doors on a better solution.

I'm not looking for suggestions on what the damage types should be (I already have a pretty good fit), rather, I'm looking for an alternate system.

Creating damage types on-the-fly is no solution either because, earlier content will find itself unable to properly ward itself against these new damage types and revisiting earlier content will be impossible (as there is a physical component to the game).

In essence, it feels like the way Mark Rosewater explains why a game such as Magic the Gathering can never have a 6th color, because of the conscious decision to make 5 colors a finite reality. I'm wondering if I can forfeit making this design decision to keep my design more flexibility.

Have you faced similar challenges? How have / would you address this?

So what's the goal/use ?

Some rock-paper-scissor effect to increase/decrease damage ?

in M:tG they just make sure they make plenty of usefull cards for every color.

One possibility is to have a system under the system. For example don't literally have fire resistance, ice resistance etc, have lower-level attributes such as reactivity, stable heat range, hardness, flexibility, etc. Then damage types can be created as a combination of the lower-level damage types.

However as others have said, sometimes keeping it simple is best because the player can understand the system better. If the system goes too far against the player's expectations, then the player will think it's broken or too difficult.

I will answer in a philosophical way. You don't want your design to be flexible Even if you can, you want to impose restrictions on yourself. Flexibility is bad, veeery bad. Even evil I would say

I agree, most of the time. I'm referring to a very pragmatic type of flexibility. One that could see use very soon, and not just a 'I can't shake this feeling that I might need this'.

So what's the goal/use ?
Some rock-paper-scissor effect to increase/decrease damage ?

Pretty much. I'm currently using 6 different types of attacks (3 physical, 3 magical). Each of them have a 'comfortable' set of purpose. One of the physical attack types is mostly used for ranged attacks, so whenever I make a ranged unit that uses a different attribute instead, or whenever I use a melee unit that uses the ranged attack type, I'm essentially breaking the mold a bit, on purpose, to give a unit some usefulness beyond just being 'stronger'.

Additionally, some effects may trigger based on the type of damage that hits a unit (burning for example), but these are generally described as a standalone ability. Still, it helps to have a damage type that is on-theme (but I could have gone without it).

in M:tG they just make sure they make plenty of usefull cards for every color.

It is a bit more complex than that. Basically, they have a color wheel that determines what color an effect should be. In addition, they have a few 'splashes' which are effects that fall between 2 or more colors. Over time, they have shaped the color wheel into a sturdy construct where everything is defined. If they were to add a new color, it would not have a personality, nor any effect per se. More importantly though, it wouldn't have enough cards to compete with the others and wouldn't be drawn at the back of their cards (one of the many issues when dealing with physical components).

My issue is similar in that the damage types will have a dedicated field (especially its corresponding armor values), and I can't start fitting 7 icons on newer units if I had 6 icons before.

When there are too many damage types or elements they just become meaningless.

I agree. I feel that 6 different types is actually a lot, but I can't help thinking I'll need another one in the near future which I can't possibly guess currently. Much like MTG here, it was impossible for Richard Garfield not to know that not all of the world could be explained within 5 colors (or that, perhaps, it was mechanically relevant, but broke its own theme).

One possibility is to have a system under the system. For example don't literally have fire resistance, ice resistance etc, have lower-level attributes such as reactivity, stable heat range, hardness, flexibility, etc. Then damage types can be created as a combination of the lower-level damage types.

I understand. If it was a PC-only game, I would've gone that way, despite how complex a system that could be. However, given the physical components, I can't afford to have this many stats.

However as others have said, sometimes keeping it simple is best because the player can understand the system better. If the system goes too far against the player's expectations, then the player will think it's broken or too difficult.

Back to square 1: make the design decision and it will bite you in the ass later.

That's what I've been doing for the longest time, but I feel there's a 'better way' that I'm missing here...

None of the expansions add new colours but they all have a distinct theme and introduce new mechanics.

This is true, because they use law-breaking abilities that re-employ the same underlying mechanics.

For example, death touch simply turns a source of damage into a lethal source of damage.

However, a mana symbol is a constant decided from inception.

So you can add new abilities

Yes, I can create new abilities on the fly because they exploit existing systems without requiring these systems to change.

and damage types

No, I cannot create new damage types because they exploit a system but define new guidelines for pre-existing content.

For example, if I create a new damage type, say 'bubblelicious', any previously existing unit would therefore have a '0' armor value against bubblelicious by default, making any new critter with the 'bubblelicious' damage type essentially deals 'all of its damage' to any unit created before it.

This is giving newer units too much of an advantage over previously generated content.

In Magic: The Gathering, they call this the power curve. In any collectible game, breaking the power curve essentially tells your long-time players to get lost. This results in bad PR/Marketing and is prone to kill your game.

As Mark Rosewater pointed out in the 2000s, the Champions of Kamigawa block (which received poor response from player, but might very well have saved Magic altogether) was built essentially as a response to the creeping power curve. Up to that point, each set introduced cards that would simply annihilate any previous set too easily (think the original Mirrodin or Urza blocks). Champions of Kamigawa was created weak on purpose (save for the unfortunate Jitte) both to support limited play all the while slowing down the power creep, giving players the impression that their powerful cards bought last year had a lasting appeal (instead of being overtaken within 6 months).

Because my system relies on very few hitpoints and a lot of armor (you could say I'm enforcing rock-paper-scissor in a way), I can't afford to have this form of power curve.

As a result, unless I can devise a system that accommodates new damage types without breaking the curve, I can't have new damage types in the future, and this is a door I'm trying not to close.

I've never faced this problem and I see how there could be benefits to limiting how much you can expand. But if the idea is to try and accommodate expansion, rather than have existing characters and entities respond to the characteristics of a new damage type, couldn't you have a new damage type respond to the attributes of characters and entities?

Maybe, Bubblicious attacks reduce the protectiveness of metal based armors by 25% except when worn by air elemental creatures on Thursdays in which case it becomes 35%. Wooden or other plant based creatures are immune to all attacks of this type. Otherwise, standard rules for attack power and efficacy apply.

I think it depends on how you use your types/keywords. If you do it in such a way that Attack Type: Fire will only do extra/less damage against a unit if the target has either Vulnerability: Fire or Resistance: Fire, you essentially close the loop. If you suddenly add Attack Type: Electricity, yes, it won't really do anything, but it also won't be overpowered. It will be more work, in that if you add that attack type, you need to add the corresponding Resistance/Vulnerabilities. Though honestly, in a video game, it shouldn't be that hard to update that kind of a system and add/remove keywords. It's not like a card game where you have to reprint cards and end up with multiple versions floating around.