A few comments occur to me.
1) I do not think the outcome described at by IDontThinkSo is that ahistorical. In reality the Germans kept pouring troops into Stalingrad in an attempt to subdue the entire city. They came very close, but were left sitting in a salient and had depleted both shoulders to the point that the Soviets were able to counter attack and isolate the German Sixth army, which was forced to surrender a few weeks later. Hitler was so focused on achieving the geographic objective (Stalingrad) that he left his army depleted and vulnerable to destruction, which sounds rather like the situation in the game described by IDontThinkSo.
2) In the old playability vs. simulation debate Unity of Command has gone more toward playability than simulation. It's a relatively low complexity game though it seems to me to have more simulation value than most other games in this category (e.g. Strategic Command 2, Commander: Europe at War, Panzer General). For instance it does a very elegant job of modeling supply. The ability to combine movement and combat in any order, along with the overrun feature, the ability to return to a unit after moving a second unit, and the ability of friendly units to negate zones of control, act together to give a pretty good simulation of combined arms and the "rhythms" of World War II combat (sometimes fast and fluid, other times a bloody slog).
Overall then I think this game is outstanding as it is, a very enjoyable, playable game of relatively short, challenging scenarios with reasonable simulation value. I think Unity of Command is best left pretty much as it is, perhaps with some tweaks. If you start adding complexity, or creating much longer / larger scenarios, in an attempt to remake it into something different you run the risk of losing what you have. When you want more complexity and realism and larger "campaign" scenarios (as I often do) there are other outstanding games such as "War in the East", SSG's Decisive Battles series, and Decisive Campaigns (who will soon be releasing a game "Case Blue" covering the Stalingrad campaign) which are available to fill that niche.
3) That said, I agree with gdrover that the criteria for determining the level of victory appear to be too simplistic. My suggestion would be to take the score that the game currently calculates, apply scenario specific modifiers for casualties inflicted and for casualties sustained, and use that modified score to determine the level of victory.
You read my mind jeffreys-
1) With the historical starting forces, setup and goals for the Axis, and considering the actual chance of a successful outcome (slim to none), what is called "gamey" seems to me to be more the historical lack of wisdom of the German high command in thinking that taking Stalingrad at any cost was worth it. You, as commander of Army Group South (who was it at this time, Von Runstadt?) are tasked to take Stalingrad, do or die. What happens afterward is of no consequence, as it is a forgone conclusion.
2) I have played just about every method of wargame simulation - RTS, WEGO, IGOUGO, Impulse (round by round each player taking one action), etc. I actually do not believe it matters to realisim that much. What does matter is the more are the attributes, opportunities and constraints the system establishes have the combatants perform historically. Here with no stacking, single attacks, overruns, ability to partial move, attack, move other units, come back to unit not completed their move, and single action to attack that can be forgone for increase mobility, steps, attachments, air, bridging, logistics, etc, the game is a masterpiece of balanced "design for effect".
3) Victory conditions are, in my opinion, the weakest link in UoC. I would have liked double the objectives with taking enough to win (say 4 out of 7), and extra credit for taking more as an alternative. I also would like to have seen more rail depicted, but witht the player needing to expend assets to "repair" rail lines, so players had more flexibility to choose direction of attack.