First of all, I want to say that I have been one of Derek Paxton's biggest fans ever since Fall From Heaven came out. That mod was genius. Derek basically took an existing game and made a mod so good that playing the mod was better than the strategy game itself. And not just ANY game - this was a SID MEIER game. So Derek, if you're reading this, I just want to say that I'm a great admirer of your work and although Stardock's launch of Elemental was suboptimal, they certainly picked the right person to solve their issues.
That said, this isn't exclusively a fan-praise post. I'm writing about a significant flaw in Elemental, namely, the game's battle system. I'm smart but not a tactical genius, yet somehow I find myself routinely defeating armies that are alledgedly three times my power or more. This is a problem. A more significant problem is that battle isn't very INTERESTING - it's basically slight variations of "march your forces up to each other and have them attack as much as possible, first strike has a a mild advantage." Yawn.
The following are my suggestions for improving tactical combat significantly to make it more strategically challenging, but most importantly, to make it more FUN. Having been a programmer myself at one point, I'm cognizant that there are only a limited amount of resources that you can sink into any given problem, so I've tried to limit my suggestions to changes that would be easy to implement.
1) Interesting Terrain
Part of the problem is that terrain in Elemental doesn't DO anything other than give the occasional +10% modifier. And even then, those terrain pieces are always off to the side, never in the center of the battlefield so the marginal advantage that they give is offset by the fact that you have to pursue the terrain. Why? Making terrain more interesting will give players much more interesting tactical decisions to pursue. The trick is that you have to throw terrain right in the thick of the battlefield, where it actively effects the combat. You also have to make the type of terrain dependent on the tiles that the two units are in before combat occurs.
For example, suppose that moving through swamp terrain costs double the action point expenditure. Armies with lots of archer units will favor engaging in swamp terrain, since they can turn enemies into pincushions while said enemies are slogging through the swamp to get them.
Likewise, suppose forest terrain blocks arrow, giving a signicant penalty to hit a unit in that terrain. That would result in infantry units favoring forest terrain for the cover it provides, and they would take more circuitous routes (both on the overland and on the tactical map) to stay in the trees.
Hill terrain could have areas of higher ground, which confer an attack bonus to units that have higher elevation than their opponents. Furthermore, going uphill costs an extra action point. Suddenly, hill become great places for defneders to fortify and establish bottlenecks.
This are all interesting terrain features that would lend a much more strategic element to tactical combat, and would also force players to consider the overland map when planning their battles. But why stop there? You could throw in interesting random elements on the map as well. For example:
A carnivorous plant that is immobile on the tactical map, but damages any unit that starts in a square adjacent to it.
Boulders and tree trunks can provide cover from ranged attacks, giving units that hide behind them a significant bonus to defense against ranged weapons.
A healing monolith that restores 1 hp per turn to units that stand next to it.
A river that units cannot pass through except in shallow areas where the river may be forded.
Not only would these elements all make tactical battles more interesting, but they lead to all sorts of interesting possibilities for new spells. For example, an Earth spell like "Boulder wall" (summons a row of boulders in an area of your choice) or a Water spell called "Liquid Terrain" (turns an area of the tactical map into swamp terrain).
2) Interesting weapons and armor
Lets face it, weapons as they are now written don't give you many choice. There's damage, and then there's more damage. Some give you a bit more combat speed, but that's a questionable choice since every attack you make results in you taking damage as well. What you need to do is make weapons with interesting bonuses and penalties, so that arming your troops will no longer be a simple cost-benefit tradeoff but a legitimate tactical choice. For example, here are some suggestions that I have to make weapons more interesting.
Reach weapons (such as halberds, polearms, and longspears) allow you to attack enemies from one or two squares away. Furthermore, if you are two squares away, enemies can't counterattack unless they also have a reach weapon. This might at first seem disadvantageous when compared to ranged weapons but since melee weapons only cost 1 action point to attack (rather than 2 for ranged attacks) then reach weapons have an extremely effective killzone unless an enemy is fast enough to dash in close.
Impact weapons (such as mauls, clubs, etc) not only inflict damage but knock the enemy back a square if the attack inflicts enough damage. Having the ability to inflict forced movement enables you to make more more interesting use of battlefield terrain. For example, if the carnivorous plant from my terrain example were used, you could push an enemy next to it, causing them to take damage.
Threatening weapons are quick weapons (like daggers, rapiers, shortswords) which enable you to take advantage of openings in combat. A unit with a threatening weapon "threatens" any square that it can attack in melee (typically just adjacent squares). If an enemy unit moves from one threatened square to another threatened square, then the unit which is threatening the second square inflicts a small amount of automatic damage on the enemy unit. This allows you to set up chokepoints even on the tactical map, stopping enemy units from getting to your soft targets.
Having these weapons would result in your choice of equipment being a much greater part of the battle. Furthermore, monsters could be made more unique and interesting by giving them or more multiple weapon qualities. For example, an giant could have both the "impact" and "reach" qualities, allowing it to push units around the battlefield from two square away. A hydra could have both the "threatening" and "reach qualities", making it a horrific threat to melee combatants.
3) A AI that uses strategy in tactical combat
Right now, the AI is woefully unprepared to deal with human opponents. While it's commendable that the AI targets weak units, it never calculates combat speed to see if it can CATCH the weak units. This results in easy wins for human players simply by running a weak but quick unit in circles while archers do the rest of the work.
What you need to do is have an AI that thinks in more general "big picture" terms and asks general questions to determine a specific strategy for that battle. For example:
"Do we have more artillery?"
Compare ranged attack strength of AI army to ranged attack strength of enemy army. If it is greater, then instead of engaging enemy units, have them go to choke points and only attack enemies that try to get by. If it is lesser, than try to engage in melee as quickly as possible, using areas of cover as a path to reach the enemy.
"What dangerous units does the enemy have and how can I best respond?"
Dangerous units are units that have a very high attack strength. If they have low defenses and low movement, then have archers target them while spreading your melee units one space apart and going after them (this stops them from using the "run in circles" technique. If they have high defenses and low movement, have your units focus on low defenses targets first, and if the high attack strength unit gets close, have your units use all of this movement to run away. This results in the enemy wasting its most dangerous unit chasing yours around the battlefield.
"Do I have regenerating creatures?"
If so, use regenerating creatures as melee combatants. Send them to the front line to fight until they reach half their hit points, then have them run to cover as quickly as possible and move through cover to the rear lines until their wounds heal. Then send them forward again. Repeat as needed.
There are also a few strategies that should be used for overland AI. For example, I always send a few low-level heroes in combat with my really tough armies, even if the heroes don't have the weapons or armor to engage the enemy. In combat, I simply have those heroes hide in the back while my army destroys the enemy. This lets the heroes grab a share of the XP and level up quickly, so they are always far more powerful than AI heroes. Well, why not have the AI use a similar strategy. It would be really invaluable, especially in the early game.
Anyway, these are just a few general ideas for making tactical combat more interesting. Elemental is a great game in principle, but boring battles really detract a lot from the other great elements of the game, and that making some significant changes to make combat more interesting will result in a lot of happiness among your fan base. 