The only argument you guys are making against 1UPT is that "Civ is a strategy game, not a tactical game!" It's great you feel that way, but that doesn't explain WHY SoD is better than 1 UPT. You need to explain why one is better than the other in Civ without clinging to the strategy vs. tactics distinction.
First, any system which does not strictly limit the number of units in a tile/square/hex/whatever per definition is a "stack system".
Although the combat system of Civ5 was announced to be 1upt, this is not true. We have stacking airplanes, we have stacking generals (which for this purpose have been transfered into 'civilian units') and we have stacking of military and civilian units in general.
This was obviously caused by the simple fact that 1upt just does not work. At least not on typical civ-game maps.
The sad truth is that you don't have enough space. Furthermore, without stacking there won't be "capturing" of settlers and workers, as this is reserved for melee type units which then have to occupy the same field as the potential victim.
1upt does not work for a civ game, and the evidence is Civ5. It runs by a xupt system, with x typically equaling "1" or "2".
Second, due to the attempts of making the x in xupt as small as possible, the workload for the core engine to calculate movements and positions was drastically increased.
The result is that organizing the movement of a very small number of units (in comparison to say Civ4) takes considerably more time, thus resulting in longer waiting for the player.
Furthermore, misplacement of only one unit easily results in unavoidable misplacing of other units as a result of some kind of "chain reaction".
Say, you have six units A, B, C, D, E, F.
For the best battle result, unit C has to be put into a certain hex.
Now, due to movement restrictions (and I am not talking about weak programming here, this is just an additional problem) it may happen that unit C is to far away to occupy that hex. Therefore, the AI moves unit A into that hex.
At the next turn, unit C may be in range, but cannot be moved into that hex, as it is still occupied by A. But A would belong behind C, which now isn't possible. Therefore, F is positioned behind A. But F would have been needed to "guard" E, a task which now is assigned to B, which in turn should have been positioned for flanking.
You get the picture.
The stack in contrast allowed the AI to group a given number of units and to move them all at once, avoiding pathfinding and positioning problems.
As long as so-called AI's are weaker than the human brain (and this will be the case at least for the next 20 years) any additional load in terms of computing necessities actually reduces either the competitiveness of the AI or enhances the computing time significantly.
In Civ5, both happens.
Whether you want to call it strategy or tactics, 1 UPT requires much more thinking while at war instead of just dragging a massive SoD from one city to the next. How was that ever fun to any of you? I played the hell out of Civ IV, but war way too simple and repetitive. That's why I think 1 UPT is superior and overall more fun than SoD.
This so-called "argument" doesn't become more true by repetition.
Battles in Civ5 (as in all civ games so far) are fought over the control of cities.
So, the target hex is just one hex, that way leading to the fact that in almost any case there isn't any point in having more than ~6 attackers (+/- 2).
The defender typically needs/has even less.
So, a group of 6 units is approaching the target.
Due to questionable design decisions, placing attackers onto open terrain typically is a bad idea, at least as long as there are some defenders around which could perform counter-attacks.
The result is that the attackers will move through rough terrain, limiting the number of feasible hex to be used for approach.
Ranged units should be positioned in the rear, units with strong values in the front.
If the defender has only a few or no defending units at all, it becomes even more easy.
I have to admit that I am quite astonished that people call this "requiring more thinking".
The way in which a target has to be approached should be a no-brainer for a decent human player. Only exception would be to be out-classed by the enemy units. But then typically you would like to avoid the attack anyway.
Now, let's have a look at the "stack attack".
As long as attacking and defending stacks are of similar strength (in terms of numbers and quality), attacking a city does require some thinking, too.
You may want to sacrifice some weaker units (less promotions, weaker class, whatever) just to soften up the defense.
Sometimes, a battle between the attacking and the defending unit just has a completely unexpected result; the defender may have taken no damage at all, or has been redlined, or whatever.
The effect of this is that now not the second attacker (according to your initial plans) would be the best option, but attacker #5. And once again, combat results may cause for changing plans.
A good attacking (and counter-attacking) requirese quite some thinking with stacks, too. At least as long as you would like to have next to optimal results, meaning to lose only that many units, to take only that much damage, to be enabled to move on with your attack force.
In total: the way in which "battles" are resolved in Civ5 doesn't require considerably more effort from the human player.
For the human player, any of both systems can be easy or difficult, based on the circumstances.
Yet, for the AI the "1upt" makes it *much* harder, effectively making battling much easier for the human, as there isn't a worthy opponent.
What was "gained" by switching from stacks to "1upt" was not more strategy, but more tedious manouvring without benefit; and I haven't even mentioned the infamous blocking due to neutral units somewhere on the way.