I am concerned however, how will they make large army movement possible without stacking, you must be able to move many units at once to speed up movement of large masses of units. It would be horrible to have to move every single unit one by one in a large army.
Ok, here´s how the old classic Panzer General does battle:
Let´s take two infantry units against each other. The defender has 10 hitpoints and the attacker is an experienced unit which have the possibility of OVERSTRENGTHING, so let´s say the attacker has 15 hitpoints. The attacker throws its hitpoints against the defender and does damage according to its attack values, infantry might have attack values like SOFT ATTACK 5 and HARD ATTACK (against tanks) 2, AIR ATTACK 1 and so forth. Of course an overstrength unit does more damage than a normal unit, it has more men and more firepower, logical.
A more advanced infantry unit might have more firepower so it has higher attack values and so on.
So the attacker attacks and defender defends, the defender has level 5 entrenchment, it has spent 5 turns fortifying, so unless the attacking unit is an engineer unit which bypasses entrenchment, the result will be a costly combat for the attacker and more so if the high entrenchment triggers a "combat event" like "tenacious defence" or "ferocious defence" which increase defence effectiveness highly resulting in even more casualties for the attacker.
Of course there could be offensive combat events as well, giving bonuses for the attacker, like "fire support" or "shock" etc. Tanks have the capability of overrunning weak defenders, they simply blow them away and continue movement.
When defending artillery units have the capability of SUPPORTING FIRE, they fire in support of defence when in range, a powerful artillery is key in both attack and defence. Attackers use artillery with longer range to destroy the defenders artillery and to reduce entrenchment, more entrenchment weakens artillery´s effect, so it might take several turns of bombardment to weaken a well entrenched foe.
In People´s General it was possible to do air support missions, you could place a fighter-bomber over an area and it would fire in support of defence within nearby hexes, air strikes could also be used to bombard defenders, although any nearby anti-aircraft units would fire on an attacking plane making airstrikes very risky when the enemys anti aircraft units had not yet been destroyed by artillery or by air defence suppression air missions
All in all, the PG series had a good combat system, they used multiple values, but as I´ve said all the same effects can be done with one attack/defence value which uses modifiers to simulate different effects. Hitpoints would be good to have I think. You could have units which have lots of men but little firepower and small units with few men but immense firepower.
You could have combats with 15 hitpoints versus 10 hitpoints, even if 10 hitpoints has more attack strength and wins all 10 dice throws and kills 10 points of the enemy, they would still have 5 hitpoints left. every strength point could both inflict damage and parry damage, but that would lead to an automatic win if you have more strength points, 5 str vs 10 str, 5points are blocked but 5 get through and do damage
Or alternately you would have hitpoints and strength points, in combat dice would be thrown as many times as there are strength points, who gets a higher score wins and does damage, example str 5 vs. str 10, 5 throws result in 500 and 10 throws result in 10, damage done to loser according to damage value, some units could have a higher damage value, doing more damage when they do, FIREPOWER value
Also, the combat dice could be "weighed" in some cases, for example tanks could get a modifier giving them more points per every dice throw making even a hitpoint 1 out of 10 tank lethal. Weighing the dice could simulate tactics, tactically superior units are more likely to win dice throws than disorganized barbarians, like Roman legions vs. barbarians.
Let´s hope we get a good combat system for Civ 5.
Cheers!
PS.
- "every round is a hit" , I think that would be unrealistic, my idea is to simulate real world dynamics with a mathematical model like modifiers and values, every value and mechanic MUST have a real world foundation.
In battle not every bullet hits, muskets were very inaccurate, archery was difficult etc. Ranged combat should have accuracy as a value.
First strikes simulate ranged combat on a battlefield, for example two units start at 1000 meters range on a field, other one is swordsmen other one is archers, they close each other, at some hundred meters distance the archers fire a volley (1 first strike), they can fire multiple volleys (drill promotions, can fire quicker, more first strikes) as the swordsmen close the distance to melee range, when the melee begins archers should be vulnerable, they should be easy pickings for swordsmen on open field (-50% defence) so in open field, unless archers do away with the enemy with first strikes, they are gone.
In musket era it is the same thing, two musketeers both have first strikes, both fire once or more at each other before melee begins (bayonet attack), well drilled musketeers could fire more volleys quicker than untrained ones. (more first strikes)
When two units with first strikes meet, what determines who fires first? Should there be an INITIATIVE value, to determine that?
So definetely keep first strikes they simulate ranged fire on the battlefield when two units meet.
Paradox uses a system called "combat phases" to simulate ranged combat in europa universalis series which simulates history in the time period 1419-1820 including both medieval era and musket era combat. In this system you first have FIRE PHASE (ranged combat) followed by MELEE PHASE, basically the same system, ranged units fire first and then the melee begins.
The question is how many times can an archer unit or a musketeer unit fire BEFORE the melee begins? Well trained archers can fire multiple times and the amount of time they have depends on how far a bow can shoot (500 meters? or less?) and how many shots you can shoot in that time as the melee units close in (probably running and screaming wildly so that lessens the time you have), so how many seconds do you have before the melee units have run the distance from about 500 meters (maximum range of the bow) or less to you, how good were you in school gymnastics when they had running, I sucked but still 100 meters doesn´t take that long even if you walk. (Roman legions walked because they were keeping formation and because they had large shields to take cover behind, I presume you CAN run even with an armour shirt on, it´s just harder and heavier.)
The same problem was with muskets, you could not shoot accurately beyond a few hundred meters, so a melee unit could easily charge the distance and close in quickly, the musketeers might have time for ONE volley but no more, UNLESS they were using advanced tactics like PLATOON FIRING in which not all musketeers would discharge their weapons at once but one out of maybe three rows would fire and then move to reload while other rows fired, this kind of quicker firing could do more damage.
So in my opinion the combat system must reflect real world dynamics in its modeling, every mechanic MUST be based on a real foundation, that is what are you trying to simulate with the mechanic, give examples?
For example, what does "damage absorbtion" mean to simulate, what real situation?
Let's go back to discussing the interesting civ5 combat mechanics and how they might work. 
