Criticism of Traditional Macromanagement

Earlier I alluded to problems in the traditional macromanagement approach in RTS games. As previously mentioned, I'm using Starcraft: Brood War as a case study of the archetypal real-time strategy game.

In a typical Starcraft match, you make an educated guess about which branch of the tech tree your enemy will develop, and you race for a tech that you think will defeat him.  From there, you build the units you researched and attack.  Of course it's much more difficult than that because of economic constraints.


One could argue that my analysis is flawed because it's often possible to change the course of your tech mid-game. Still, that doesn't change the basic modus operandi;  in Starcraft, you race for tech and for numbers.

Why is this so?  What are the laws of gameplay that ultimately reduce to carefully timed build orders?  In our analysis, there are two:

1. The existence of a single, predictable win condition
2. Static properties of units

The first so-called gameplay law should be fairly obvious.  There is only one condition whose fulfillment can win a Starcraft match -- you must destroy all enemy buildings.  Some games attempt to prevent gameplay stagnation by allowing players to select win conditions.  Unfortunately, this does nothing to solve the problem, as once the game is started there is still only one objective.  Ultimately, the game will come down to an unimaginative battle.

The second law is perhaps less obvious or more prone to confusion in the minds of developers.  Static properties of units does not imply that the competencies of a unit do not change during a game.  It simply means that they do not change dynamically in response to their environment, nor in real-time.  You can can upgrade units in Starcraft, but one marine and one tank working in tandem will always have the attributes of sum of both units.  This makes any form of unit interaction superficial, and this is where the gestalt principle comes into play.

If you want a player's strategy to be adaptable during the course of a game, his units strengths and weaknesses must depend on the gamestate.  A simple model for this would be the sum of the units properties, plus a cooperation bonus.  Obviously this isn't the most engaging solution, but hopefully it is sufficiently to illustrate the criticism.


Before concluding, I should mention that there is one potential pitfall that can come up as a result of unit interaction bonuses. If left unconstrained, group strength can grow exponentially with each additional unit. This is obviously bad, but what about linear progression? This could make early-game encounters particularly weighted. Perhaps it's best to constrain these bonuses to interaction between unit types?

Thoughts? Let us know in the comments!