Thought I'd add in my lot to the discussion. I've actually been planning to modify the current combat mechanics within the current available rule set. Mostly I've been testing different stats but the general gist of the ideas I've been having are below.

Note: All of what is written below is possible by modding the current game (no code changes). I just wanted to show what is salvageable from the current system that can be tried out before any major combat system changes are done.

As everyone keeps saying the outcome of a basic attack is too random. Your attack roll is also your damage roll and your defence roll is also your damage resistance roll. My idea is to bring in two under used stats to make outcomes a little more reliable. The first thing is to rename the stats to what they actually will be so:

Attack - Rename to Damage

This is actually the damage stat. The first thing to do would be to reduce the range of damage so that it never gets beyond say 1-12.

Defence - Rename to Resistance/Soak (something like that)

This is the damage resistance stat. The final values for this stat should end up within the same range of number as the Attack(Damage) stat, maybe higher (as a high skilled attacking unit can bypass armour - see below).

Dodge - Can stay the same or be renamed to Defense Skill

This is your defensive skill and will be the sole stat the determines if you get hit or not. All units would have a base dodge of about 20% and gaining experience would raise this amount higher. Champions would either have a higher starting base or the possibility of a trait that increases the base dodge amount. Likewise dodge will be unlocked so that it can increase when leveling up. Note: the dodge skill currently operates between 0 and 1000 for 0% to 100% dodge, which makes the dodge on shields pointless right now.

Daring - Rename to Attack

This stat is currently only used for one trait. However it can be usurped to become your attack skill. Assuming your target does not dodge your attack then your Daring(Attack) skill will come into effect. This stat gives a chance to ignore Defence(resistance). All units would have a base dodge of about 20% and gaining experience would raise this amount higher. Champions would either have a higher starting base or the possibility of a trait that increases the base dodge amount. Likewise daring will be unlocked so that it can increase when leveling up.

All weapons and armours would then have to be re-balanced to take into account the way the stats are now focused. For example maces/hammers would likely have a much lower potential damage but add to your Daring(attack) as blunt weapons tend not to be affected as much by armour as cutting weapons. Heavier armours would add to your Dodge(defence) like shields currently do to simulate more blows being deflected, while adding in some other penalties like slower movement (not combat speed) and possibly lower Attack(Damage).

I would then probably increase the base hit points of units in general to maybe 10 (and remove HP scaling with experience as this will be covered with Dodge(Defence) and Daring(Attack)). Which will prevent lower damage weapons from getting a lucky one hit kill, larger two handed weapons still could.

Awesome this for sure.

Nothing beats the old school computer game tbs combat like Pools of Radiance.  I always preferred Pools of Darkness myself as having higher levels gave more abilities to play around with.

The resistance should be seperated into different categories (Slashing, Piercing, Bludgeoning, Earth, Fire, Air and Water) to give armors and creatures more options. Some armors and creatures should have a negative resistance, that increases the amount of damage (Light Plate Armor should have a negative air resistance, Fire Elementals should have a negative water resistance).

A Combat Speed of 6 increases the number of actions to 7, but i think it should increase the number of actions only to 6 (1, 2, 5, 6, 9, 10) to match the other Combat Speeds.

There are as many different ideas as how tactical combat should be implemented as there are stars in the sky. I'm mostly behind the second poster who said he liked the system as is. i.e. I also prefer simplified combat, and I think that's what the designer wanted from the getgo which is great from my POV. However, there are some problems with the combat system as it is now that I'm seeing which need to be fixed.

1) 99% of the troops move two squares a turn. There needs to be more variety here. Ground units should be able to move 1 to 5 squares and mounted units in the 5 to 10 range, flying units should be more or less unlimited. To support this, the grid size needs to be expanded perhaps another 5squares around all the edges to allow additional maneuver. Also, all units should be allowed to escape or none escape. Why allow champions to escape but not 4-man archer groups?

2) The tactical AI stinks. They focus too much on champions or whoever is weakest. I've had so many battles that I've been outnumbeerd by more powerful troops only to see those enemy troops run around the battlefied trying to run down a champion or weaker melee squad only to have my archers whittle them down from 200 hit points or so down to zero. over 15/20 rounds of combat. I use my "runner" to draw monsters/enemy away from my archers so they can shoot at long range unhindered. This is just dumb. I've had powerful trolls (152 hp) start next to one of my archers (could've easily squashed it with one blow), only to see it move two squares in an attempt to get my champion leaders who were positioned in the rear, wft?? The trolls just end up getting squashed by my archers; I don't even need to move my archers, most of the time they're always safe.

3) There's not enough difference in weapons. Swords are available later than maces yet they're weaker?? .. There simply needs to be more differences in the weapons: it's simply not showing up on the battlefield. I use two weapons in this game as of now & I win each and every time: heavy mace and ceder longbow. You don't need anything else.

4) There's not enough difference in the terrain on the battlefield. Woods should have a significant effect on archers. I can kill any unit with my archers no matter where they are, this is extremely weak & bad. There should be something like a 95% defensive bonus for woods against archers. In fact, I don't see any terrain effects on attack or defense at all. It's simply not showing up in a way I can see. The only thing woods seem to do in this game is make it difficult to see which grid square units are positioned. In fact, I've lost a few units only because I actually didn't see the enemy units behind the tree. 

5) This goes along with 2. Certain AI spells are always cast on the same unit. When a troll or magic user casts confusion, it's always against one of the melee units in the front rank, never on the more lethal archers behind, not only this but they cast the same spell on the same unit turn after turn?? wtf, is there something here that I'm not seeing? I'm assuming that the spell had its effect as it said "confusion" in red letters, but frankly, some of the spells I can't tell if they work or not if they don't directly indicate "miss". Also, teleporting units should be allowed to teleport and conduct at least one attack. Those big black spiders should be able to teleport right next to a champion and get one attack on him/her, instead they just teleport around the battlefield stupidly and do nothing with this advantage, might as well just move them forward and not teleport. In general, monsters need to have more range attack, those big slow "slugger" monsters with attack in the 75 to 152 never even get near my macemen as my archers always take them out while they lumber forward at 2 squares a turn. Those that do manage to get that far usually only have 2 to 10 hit points and get smited by the first swing of my macemen who typically dish out 20 to 40 hp of damage per swing.

6) I think that the current number of units allowed on the battle should not be changed. Frankly, I don't want to see "epic" battles because I'll just spend an hour resolving one single battle which I know how it turns out before I even fight. As an analogy: this game is not designed to be an "army" level game, but more of a "company" level game, which is fine with me as I like both as a strategy game player.

Just got the game, and I gotta say that the sheer unpredictability of combat is one of the most noticable thing about this game.  I'm of the opinion that randomness shouldn't even be a factor in combat systems like Elemental's.  The sample size is low, and if you get a high roll against the right unit, you can change the course of the entire game (forget the tactical battle).  Way too much is riding on what is essentially a crap shoot, and I don't see any reason why we should even be rolling dice in the first place.  It's too early for me to weigh in on the specifics, but at very least the results shouldn't be so wildly divergent.

My personal opinion on the subject of randomness in TBS is that it's a carry-over from the traditions of table-top gaming, and it's largely unnecessary in a computer-driven game setting.  In board games, the paperwork had to be kept to a minimum (preferably eliminated altogether), which meant the game state had to be very simple.  If you can make a few generalizations to simplify it further, it's quite possible to trivialize the game.  Randomness solves this problem, preventing you from making cause-effect corralations and forcing you to factor uncertainty.  As an upside, you get the visceral feel of the dice roll, but as a downside irrational strategies can prevail over rational ones by sheer luck.  With a computer game, the machine handles all the paperwork, so you don't need to rely on randomness for depth; you can create very complex game-states, but the human can still comfortably work with them because he doesn't need to crunch or record the numbers himself.

I disagree and believe the opposite is true; that less randomness gives the player more control, more involvement, and better rewards skill.

The problem if unit Y can beat unit X is that there's a 20% chance (or whatever it is) that one player will be rewarded for making a bad decision and the other will be punished for making a good decision.  This doesn't reward skill or increase player involvement.  It gets worse if you have 5 unit X's against 5 unit Y's.  Now there's a 66% chance that you lose at least one of these battles, which restricts your ability to enact more elaborate tactics or set ups.  In a game with larger numbers of units, that's okay since you have some redundancy, but in Elemental we don't have that.

First of all, I don't really understand why Elemental tries to simplify combat mechanics (one number for attack chance and damage). Is it to make the AI better understand the mechanics?

Generally, pen and paper games must simplify, because it's not fun to calculate hundred numbers for each sword swing. However, computers are good in number crunching, so computer games may benefit from more detailed models of gameplay mechanics - like combat.

Let's look at Dominions 3 combat system a bit - I don't know if many of you know it, but I wholeheartedly recommend it.

Weapons in this game have multiple statistics - damage, chance to hit (attack), defense bonus and reach. It's not so linear (biggest sword kills all), but presents interesting choice for the players. When there is a difference between weapon reaches (pike vs dagger, for example), the shorter weapon user must first successfully defend to shorten the distance. If he is unsuccessful, he suffers damage and is "repelled" back, and has to try again. This way, a formation of pikemen with lousy stats can hold at bay a pack of wolves - just because their long weapons offer them safety. Against a squad of mediocre, but armored swordsmen, they fall easily.

Another important point is that weapons affect defense. Obviously, a slender, renaissance sword with elaborate guard is better for parrying than a heavy, but unbalanced club. Should safety be preferred to ability to inflict damage? Here, units take different roles - one with better defense holds the enemy at bay, while other attacks at range, or attempts flanking.

Lastly, the difference between dodging or parrying the blow (defence), and absorbing the blow makes the system more varied and interesting. A heavy club may never hit an agile opponent, but if he tires down and his high defense stat plummets, first hit will will send him across the floor. A heavy armor may be impenetrable to normal blows, but a single hit with a magic wand may turn its user to stone.

Dominions 3 also has a great synergy of spells and conventional weapons. For example, Yomon has weak, but cheap archers that are nearly useless against armored troops, their arrows bouncing off harmlessly. A clever player will prioritize research of the "Fire weapons" spell, and because armor is only half as effective against flaming weapons , these weak archers suddenly become quite deadly. Undead player can cast "unnatural aging" global spell, and enjoy watching the enemy heroes wither and die of old age in a few turns, while being unaffected himself (long-lived Rylegh are another matter, though). Abysian troops from the underworld have infravision, so the global spell Eternal Darkness gives them considerable advantage.

And so on, the examples are numerous.

I am not saying Dominions 3 is better game, but some of its parts definitely work splendindly, especially the said combat system. There is a lot of detail hidden - the player may just read the result of a battle, watch it casually, or examine the log for detailed explanation of what happened ("oh, my giant got paralysed in the first round, and my footmen were decimated by arrows. Maybe next time, I will forge that protective ring and hire troops with shields").

An ambigious system where there is no clear "better or worse", but forces player to make interesting choices usually contributes to an interesting and immersive gameplay. Creating a competetive AI for such a game is usually quite impossible, though.

Well, I'm not against the epic unit number like this. But if it is implemented by the Dev, please make an option that give a single unit with the number of soldiers in it, just like in Heroes Might & Magic. So, the high end PC owner will able to play Massive version of battle, while the owner of Low end PC still able to play the game without to worry about the RAM Problem.

Really, I'm NOT proclaiming for You. I'm proclaiming this for Those Who Want It, which are counted in the other two threads going about it. You can't say there aren't a lot of people who want it, because there ARE a LOT of people who want it. If you like, feel free to count those who want it in the Two Threads going about it, please. Frogboy asked for input here, and input is what I'm giving. It's the first time I've replied to the thread so I think I have the right to say it when so many Other People (not you of course) want it and want them to stick to it even if it is "hard to balance". Something being hard to do isn't an excuse in my book not to do it. If you don't do something because it's hard to do, what does that make you?

Please, don't tell me what I can and can't suggest when You weren't one of the people who wanted the suggestion made. We all have our own opinions my friend, and we're all free to speak them. Would it be cool with you if I asked you not to say something just because I didn't agree with it? I don't think you'd like that much, and my entire argument, along with everyone else's was kept in the threads about it. If the head honcho asks for input, I'll give any input I feel the game needs, just as I expect you or anyone else to, ok

I like Walls in combat too. They've been planned from the start so we should see them once they get the time to implement them

I understand that. You wrote it once or twice. And your "official" Thread is ok. I have nothing against it. You want that and you fight for that. Everything ok. In this request-thread you describe what you want and everbody who is with you could say "Yes, i want it".

But other threads are started and written in another context. For example this thread. Frogboy asked for input and everbody is free to give this input, his personal ideas. So pls dont use words like "we" and "us".  That would be nice. thx.

Lots of good ideas in this thread.

My top three "easy to implement" wish list for 1.09;

1. Movement rates should differ - mounted and lightly armored units should move farther than heavily armed or unmounted units. This would have a dramatic change on the battle.
2. Bonus/Penalty for flank attacks - coupled with movement this would open up the tactical battle and give players a reason to favor mounted or lightly encumbered units.
3. Update the AI to "kill".

Next on my list for 1.13 would be;

1. Initiative based unit movement that uses Random Factor + Training (bonus for more highly trained units) + Armor (penalty for heavier armor)
2. Line of sight and Terrain effects - You can't attack what you can't see - except for area based magica attacks. Forest, hills and walls block line of sight. High terrain give line of sight bonus.
3. Tactical Actions for Non-summoned Humanoid Combat Units - i.e. ability for a unit to choose: Shield Wall, Withdraw (if Approached), Defend Position, Charge (+2 attack, -2 defense), etc. Combat units could select a tactical action each turn which would offer bonuses and penalties for that turn. The combat actions available to a unit would be based on their level of training and tech research decisions.
4. Combat specific spells whose effects are not duplicated in other spell books. - Water: Rain storm (reduces line of sight for all units), Air: Wind Storm (Affects archery attacks), Fire: Forest Fire - sets forest terrain on fire rendering them impassible for the remainder of the combat and blocking line of sight for units on the far side of the fire.