A real-time strategy game is, by definition, a game where players are forced to make strategic and tactical decisions in real time. As the game industry grows, the real-time strategy genre has narrowed its focus to a very specific type of game that does little to force players to consider an over-arching strategy as comprised by numerous tactics. Instead of allowing a player's large- and small-scale decisions to adapt and change as events in a given skirmish unfold, RTSs just make players think of resource usage (I have X, I need Y, and I get Z/minute) and basic army composition. Everything else in the span of a game flows from these two mechanics into what is, typically, one large battle near the end of a game. Relic's Company of Heroes changes this design and, as a result, makes its real-time strategy gameplay into a more dynamic and far less predictable experience that forces a player to make harder decisions more frequently.
It's a commonly-held tenet in real-time strategy games that when an enemy unit is right-clicked upon that death befalls it after it takes a certain amount of damage from units that deal a specific amount of damage every few seconds. Blizzard's Starcraft is practically built around a very definitive combat model that follows a rock-paper-scissors methodology with very consistent unit performance results. The micromanagement that occurs within battles in Starcraft has nothing to do with centering an army around a well-covered/fortified position or ensuring that when your Dragoon attacks that his bullets will hit the right part of the enemy siege tank; instead, cover is just determining if a Protoss melee unit is in range of a bunker filled Space Marines and any hit a Dragoon lands on a Space Tank will do the same amount of damage whether it hits the armor-heavy front or the weakly-covered rear.
The design team at Relic took a far different approach to the combat in Company of Heroes than any of Blizzard's efforts. Every part of the game map has a cover value attached to it that, when right-clicked upon, will serve as a hint to a squad of units as to how they should interact with their environment (ie, crouching behind a wall of sandbags or ducking under the lip of a crater). Under this design, two squads of riflemen with the exact same stats can face off and reach a dramatically different outcome depending on their cover situation. As an example, Squad A may be crouching behind two layers of sandbags (heavy cover) while Squad B attempts to take their position from an unfortified open road (no cover or, worse, negative cover). Since the only difference in this sort of encounter is each squad's probability of landing a successful shot (modified by their cover) on an enemy it is, theoretically, possible for each squad to kill each other at the same time. In practice, it may take Squad B three-to-four times as long to eliminate Squad A was it would for Squad A to wipe out Squad B.
The design becomes more complicated when tanks and troops wielding bazookas, panzerfausts, and panzerschrecks join the fray inhabited by the rifle squads above. Unlike rifle bullets, large projectiles in Company of Heroes are a very prescient danger that visibly travel across the screen and violently collide with in-game entities and structures. If a rocket launcher is fired and hits tangentially to a tank's front or side armor it will take minimal damage (or, in some cases, deflect off and hit a nearby structure). If that same rocket hits the lightly-armored rear, though, the take can sustain heavy damage along with a busted engine or armaments. And if that same rocket, or tank shell, hits the layer of sandbags that Squad A was hiding behind in the above example then a player can say goodbye to half of the squad along with the sandbags that were covering them.
While designing the game, Relic must have known the endless amount of abuse that these rockets could wreck upon map structure and players alike because they added a very heavy degree of variation in how a rocket could be launched or tank could fire. The developers of Company of Heroes completely violated the unspoken tenet of real-time strategy and, as such, when a player chooses to attack a target using his Tiger tank there is a chance that a rocket may completely miss a target and hit another enemy, fly harmlessly into the distance, or deflect off of a stray tank trap into a player-controlled building. A player can position his Tiger in such a way as to make a direct attack far more likely but there is, in essence, never a guaranteed strike from a rocket or tank.
The change from a fairly predictable combat design to a very visceral, dynamic battle engine is one that Relic handled to great effect but does such a degree of randomness in combat scenarios do anything to cheapen the "strategy" involved in the game? A fervent Starcraft or Command & Conquer player would be quick to point out that the lack of consistency from game to game would prevent a game like Company of Heroes from ever being considered for competitive play at a pro gamer level. That is a definite possibility, of course, but more realistically what Company of Heroes does is to provide a far more strategic gameplay experience as a result of the surprises that occur in the middle of the game. The game design provides the mechanisms by which a player with a less grandiose army can, by utilizing both cover and more intelligent rocket infantry positions, overcome a larger set of forces. And when such an upset can occur in the middle of a game that encourages tactics across numerous encounters it offers the chance for another reversal of fortunes later on. And that is the kind of strategy that can adapt and change over the course of a game which allows randomness in its design.