Temporally Constrained Games: Prison Equilibrium in N-Person Games with Finite Horizons
Abstract
We introduce the concept of a temporally constrained game, a class of n-person games in which each player operates within an exogenously imposed finite time horizon — a sentence — at the expiration of which performance is evaluated and payoffs are determined. This structural element, absent from classical game theory, fundamentally alters strategic behavior: as a player’s remaining sentence approaches zero, the set of available strategies contracts to a singleton, forcing convergence to a unique action regardless of the player’s preferences or rationality.
We define a new solution concept, the prison equilibrium, which differs from the Nash equilibrium in its mechanism of stability. Whereas a Nash equilibrium is stable because no player has an incentive to deviate, a prison equilibrium is stable because no player has the time to deviate. We prove three main results. Theorem 1 (Existence): every finite temporally constrained game possesses at least one prison equilibrium point, established via Kakutani’s fixed point theorem applied to time-dependent strategy correspondences. Theorem 2 (Convergence): as the remaining sentence approaches zero for all constrained players, the set of prison equilibria contracts to a unique point, yielding deterministic predictability near deadlines. Theorem 3 (Superadditivity of Nested Sentences): when multiple players’ deadlines coincide, the aggregate action intensity is superadditive — exceeding the sum of individual effects — producing what we term a wave.
The framework rests on five axioms. (A1) Time is an exogenous constant: no player can alter any player’s sentence. (A2) Greed is a positive constant: all players seek to maximize payoffs. (A3) Spatial freedom is a decreasing function of elapsed time: the effective strategy set shrinks monotonically as the deadline approaches. (A4) The sentence is the source of power: a player’s capability to act is granted by and persists only within the duration of the sentence; expiration of the sentence without renewal is equivalent to elimination from the game. (A5) Sentences are public information: all players know all players’ deadlines.
From these axioms we derive a wave function $W(t) = \sum_{i} g_i C_i / \tau_i(t)$, where $g_i$ is the greed constant, $C_i$ is the capability, and $\tau_i(t) = T_i – t$ is the remaining sentence of player $i$. This function diverges as any deadline approaches, providing a quantitative prediction of aggregate market action — volume, volatility, and directional flow — near institutional deadlines such as quarterly closes, options expirations, central bank meetings, and electoral cycles.
We further identify a distinguished class of player: the free player, defined by $T_i = \infty$. The free player faces no deadline, contributes no forced action to the wave function, and is the only player capable of exploiting the predictable waves generated by constrained players. We show that in the limit $T_i \to \infty$ for all players, the prison equilibrium reduces to the Nash equilibrium, establishing the classical theory as a special case of the temporally constrained framework.
The theory unifies three previously disconnected traditions. Adam Smith’s “invisible hand” is reinterpreted as the aggregate of forced actions by deadline-constrained agents — a hand made visible by the public nature of institutional calendars. Keynes’s government intervention is reinterpreted not as an exogenous correction by a neutral regulator but as the strategic action of yet another constrained player operating within electoral and political sentences. Nash’s equilibrium is reinterpreted as the degenerate case in which all sentences are infinite and time exerts no pressure on strategy selection.
Two interpretations of the prison equilibrium are offered, paralleling Nash’s original dual interpretation. The institutional interpretation views temporal constraints as arising from legal, regulatory, and contractual obligations — quarterly reporting requirements, fund redemption windows, central bank meeting schedules, electoral calendars — that mechanically compress strategy sets as deadlines approach. The universal interpretation views temporal constraint as an irreducible feature of all human decision-making systems: every agent with a finite lifespan, a finite mandate, or a finite resource horizon is a prisoner of time, and the resulting forced convergence of behavior is a structural property of bounded existence itself.
As an application, we solve a simplified three-player quarterly trading game involving an institutional fund manager (sentence: 90 days), a foreign institutional investor (sentence: 90 days, with an additional currency-hedge sub-sentence of 30 days), and an individual investor (free player, $T = \infty$). We derive the unique prison equilibrium near the quarterly deadline and show that the free player’s optimal strategy is to act counter-cyclically at deadline-induced waves — buying into forced selling and selling into forced buying — a result that is structurally unavailable to any constrained player.
The paper is organized as follows. Section 1 introduces temporally constrained games and motivates the departure from classical theory. Section 2 provides formal definitions and terminology. Section 3 proves the existence and convergence theorems. Section 4 analyzes nested prison structures. Section 5 defines solutions and values. Section 6 characterizes the geometrical structure of the shrinking equilibrium set. Section 7 develops dominance relations under temporal constraints. Section 8 presents the quarterly trading application. Section 9 discusses motivation, interpretation, and relation to the theories of Smith, Keynes, and Nash. Section 10 concludes.
Keywords: game theory, temporal constraint, prison equilibrium, finite horizon, institutional behavior, market microstructure, deadline effect, Nash equilibrium generalization
JEL Classification: C72, D82, G11, G14, G23
Mathematics Subject Classification: 91A10, 91A80
The paper is 28 pages. The revolution is not in the mathematics — which employs standard fixed-point and contraction mapping theorems — but in the definitions. Two definitions change the frame: temporally constrained game and prison equilibrium. One theorem changes the prediction: convergence near deadlines is deterministic. The rest follows.