Simulation Horizons, Constraints, and the Emergence of Strategic Agency

This essay proposes a boundary condition for when "optimization" becomes strategy. Many systems respond to local gradients (migration, adjustment, reflex-like control). Strategic behavior—incurring present cost to reshape future payoffs—appears only when an agent can simulate counterfactual futures far enough ahead and has sufficient control authority to act on those simulations. We formalize this with a constrained horizon model: thermodynamic and organizational overhead bound the effective planning horizon, while actuator limits and structural controllability bound intervention capacity. In this framing, equilibrium is less a timeless fixed point than a temporarily stabilized trajectory of a constrained feedback system, which can persist for long periods yet collapse abruptly when horizons shorten or constraints bind. The goal is not to "prove" a single theory of agency, but to make the domain of validity of game-theoretic explanations explicit—especially in regimes (crises, panics, coordination breakdowns) where game theory is often invoked precisely as the system exits the strategic regime.