Continuity in TB Combat - bad idea? (+associated musings)

[Section8]

(1) Any thoughts on interesting ways to handle magical combat/conflict (magic missile and fireball need not apply)? [NB, there needn't be wizards/spells, not all "characters" need be physically on the battlefield, victory needn't involve killing funny little men....]

Weather Effects:

• Flood the combat zone. Make higher ground a necessity and not just an advantage.
• Snap freeze the combat zone. Why cross a river at a chokepoint like a bridge, or worse, a ford, when you can freeze the whole thing? Maybe alter the whole gameworld dynamic so that changes of direction become hazardous due to low friction surfaces.
• Lightning Storms - Make those fuckers regret wearing/wielding all that steel.

Terraforming:

See Magic Carpet.

On Second Thought:

Introduce dynamics that force the enemy to revise their tactics completely. I'm of the opinion that any dynamic that really makes the player question whether or not to try and kill something, even though that's the objective of the game, is fucking sweet:

• Exploding corpses
• Pain/Death triggered enhancement - "strike me down and I'll become more powerful than you can possibly imagine!"
• Rust Monsters!
• Helpless Symbiotes, like the little sisters in Bioshock.

Not all directly related to magic, but came into the train of thought.

(2) Bearing in mind that I know it's a terrible idea etc. etc., any thoughts on using a physics system as a significant combat component? Emotional/implementational objections aside, what's the thinking on the unpredictability aspect? [e.g. non-physics-based: "aimed shot gives 73.2% chance of hitting for 60 damage", vs physics-based: "blowing up supporting pillar stands ??? chance of collapsing ??? onto ??? for ??? damage"]
I'm wondering if it'd be desirable to introduce such uncertainty. In one sense I like the idea - since it forces the player to strategize on a game-world rather than mathematical basis. On the other hand, I'm also aware that taking away the numbers would probably upset some people - and raise the whole player skill vs character skill issue with a "Why can't my explosives expert tell whether blowing up pillar X will bring down the building??". Would a physics system which allowed some prediction (accuracy according to character skill) be preferable?
[Yes - I'm aware that it's a terrible idea etc. etc., for a variety of reasons.]

I don't think it's a terrible idea. From the top - the uncertainty is fine. If a player wants to walk a more predictable path, they can. I see physics based damage as a higher risk gamble, with a naturally higher reward, assuming a falling chunk of concrete squashes someone flat, while that same person might be able to shrug off a few bullets.

And you're right that it forces the player to strategise in a more "natural" manner than simply playing the maths. It's also one of the emergent designs where the unpredicted failures are outweighed by the anecdotal successes. And I think anything that drives a player to anecdotalise is one of the best ways to make/help a player forget failures and shortcomings.

As for the player/character synergy, it's trivial to give the player the ability of their character. A character with super hearing ought to be able to detect "ghosted" movement a la Silent Storm, so why can't structural bearings be highlighted for characters with the skill to identify them? Of course it does require a certain impetus from the developer with what they regard as viable physics-based tactics, but it's pretty passable.

With regard to prediction based stuff, I think it's better left alone. The uncertainty has charm.