Allowing counter bombardment would balance ranged bombardment a little, but it would still remain an unbalanced game element.
When 2 stacks meet one another and one has artillery type units, then this stack can weaken the other stack and mop up with normal units. So both stacks will employ artillery type units and a bombardment-counter bombardment battle will insue. But then the stack with the most artillery type units will win this bombardment fight and the normal units can again mop up the survivors of the other stack.
Now, the biggest stack wins is a normal thing in civilization IV (as long as both stacks have an equal level of technological development and are fairly balanced in the units that they've employed). So, the change to the biggest stack of artillery type units wins isn't necessarily that bad (although a bit more boring), but the biggest problem is that the winning stack will typically have very few losses. In most models bombardment and counter bombardment are not lethal and thus the surviving stack won't lose units during the bombardment phase. And during the mop up phase by normal units, the winning side will also not suffer losses as the weakened opposing units have very low chances of victory. So such a model would allow you to defeat a somewhat smaller stack of units without serious losses which would of course lead to an unbalanced model of war.
If the counter bombardment were lethal and thus a different form of bombardment than the original bombardment, then I think it would be fair that this counter bombardment would lead to a counter-counter bombardment and so on. So instead of normal units being able to defend and engage attacks by artillery type units (like it is now in civ4), only artillery type units would be able to do so and this would lead to an all out bombardment between both stacks. It would be smart for the attacker to first perform this all out bombardment phase and only afterwards add the normal units to the stack to mop up the survivors in the enemy stack. This way, they wouldn't be wounded by the bombardment between both stacks.
I just don't see how a combat model with ranged bombardment from artillery type units cannot degrade the combat model to 'who has the most artillery units' wins and wins with low losses.
The situation of ranged artillery bombardment isn't similar to ranged bombardment by bombers for various reasons. Bombers can be counter attacked by the fighters of various cities and thus the defence against this type of attack is pretty strong. Bombers also have no realistic chance to beat fighters.
Furthermore, you can only station 8 airplanes (4 without airport) in a city which disallows the situation where you create an airforce which can be used to overpower the enemy locally.
If a similar limitation were used to the number of artillery type units in a tile, then a ranged bombardment model with artillery type units might work. But it would still be more powerful than the bomber-fighter model since the attacks of artillery aren't countered in an equally lethal way.
About realism again: The WW1 situation was brought up as a situation where artillery was dominant on the battlefield. That is probably true for this war. It was the first war where it was possible to create a continuous front line without gaps because of conscription of huge parts of the densely populated European lands and because of the power of the machine gun which could easily repel attacks from larger forces. Therefore, faster units like cavalry couldn't flank the other sides positions to attack the artillery. Strong units that could power through the front line to attack the artillery positions (tanks) only arrived very late in this war and were not very efficient machines of war yet. New units that could circumvent the front line and attack the artillery positions directly (bombers) also weren't efficient enough to heavily bomb these artillery positions. It was also the war where artillery increased its range dramatically so that it didn't have to be close to the front line.
It is true, in this war, artillery was very well protected from a counterattack from flanking units, units that could power through the front line or units that could circumvent the front line. It was a unique situation which only lasted for a decade or so. Before that time cavalry could flank the positions and attack low range artillery (cannon like artillery) directly, after that time, tanks could power through the front line to directly attack the artillery and bombers could attack them from the air. I don't think we should base a combat model of Civilization on this unique situation.
Artillery wasn't immune to counter attack during most of human history and therefore I don't think that a combat model where you can bomb the enemy during your turn where the enemy cannot counter attack these units would be a realistic combat model. In a real time combat simulation, you also don't let your units sit through a bombardment without an attempt to take out the attacking artillery units.
My preferred model was already explained in post 3 of this same thread.
What we really need is a game-within-a-game like in Master of Magic.
I also loved Master of Magic. A great game in many ways and probably a bit ahead of its time. Still, I think its combat model would be a bit too detailed for Civilization. It would start to take too much time in the industrial and modern age to fight all of those battles by hand. I like the depiction of war in this model, but I do think it should be an automatic fight scene with the option to speed up time in this fight scene or completely skip it to speed up the game.
The Master of Magic combat model also had far more variables with magic spells, healing and all kinds of special abilities of the units. It might get pretty boring fast with the simpler units of Civilization.