When the Chessboard Changes: Game Theory and Adaptive Strategy in Modern Conflict
Morgenstern and subsequently developed by John Nash, among others, is one of the most powerful tools devised for understanding strategic interaction. Its fundamental contribution lies in demonstrating that a decision cannot be evaluated in isolation: the outcome of our actions also depends on what the other party will do, on what we think they might do, and on what the other party thinks we will do.
The ongoing conflict between Iran and the United States can be interpreted through the lens of repeated games, reciprocity, reputation, the credibility of threats, and response strategies. This perspective allows us to understand an essential aspect of the conflict: each actor must continually assess not only the immediate effect of their own action, but also the response it will elicit from the adversary and the consequences for subsequent interactions.
The protracted nature of the war, however, suggests a further problem: the interaction between players not only alters their subsequent moves but can progressively transform the game itself. It is here that game theory—formalized nearly a century ago and subsequently developed into increasingly sophisticated forms—encounters the complexity of the contemporary era. Not because it loses its explanatory power, but because it must grapple with systems in which the decisions of actors—through the reactions and adaptations they generate—can help reshape interests, constraints, opportunities, and relationships, thereby transforming the conditions under which strategic interaction continues.
From Games to Complexity
Game theory allows us to model extremely sophisticated forms of strategic interaction and should not be confused with a simple game of chess. Chess, however, can help us understand—by analogy—a fundamental distinction between complication and complexity. Chess has long been one of the most widely used metaphors for strategy (the knight’s move). Two players, conflicting interests, pieces with different capabilities, complete information about their positions, and a sequence of moves in which each player tries to anticipate the other’s moves.
Chess is complicated, but not complex in the true sense of complex adaptive systems. The distinction is fundamental. A complicated system may comprise an enormous number of elements and combinations, making it extremely difficult to identify the best solution, but it retains a substantially defined structure. In chess, no matter how vast the number of possible configurations may be, certain fundamental conditions remain unchanged throughout the game: the chessboard retains the same dimensions, the pieces maintain their capabilities, the rules do not change, there are still two players, and, above all, both know what it means to win. Complexity introduces a different condition. Interactions among actors produce feedback, adaptations, and emergent effects; nonlinear relationships can amplify or attenuate the consequences of individual actions, and, above all, the behavior of the actors helps to modify the environment in which they will continue to act.
A complicated system makes it difficult to calculate a move; a complex system can alter the conditions on which that move was calculated. Strategic competition among states increasingly exhibits these characteristics. We are not simply playing a very complicated game of chess. We are dealing with interactions in which moves can metaphorically alter the chessboard, change the value of the pieces, bring in new players, transform some of the rules, and, under certain circumstances, even redefine what it means to win. We are therefore not merely facing an increase in complication: we have entered the realm of complexity. It is precisely at this point that we return to game theory. It is not a matter of abandoning it, but of understanding how to apply it when strategic interaction takes place within a complex adaptive system, in which players’ decisions help transform the environment, the incentives, and the space of possibilities for subsequent interactions. A theoretical foundation for this convergence already exists. Studies on complex adaptive systems describe systems composed of heterogeneous agents that interact, learn, and continuously modify their behavior. W. Brian Arthur has shown, in complexity economics, how agents with imperfect information explore strategies, react to collectively produced outcomes, and adapt their behavior. Other studies have linked game theory to complex adaptive systems: the former offers rigorous tools for analyzing the interdependence of decisions, while the latter allow for a more complete incorporation of learning, heterogeneity, feedback, evolution, and emergence.
When the Game Itself Changes
In the simplest representation of strategic interaction, we can imagine players with specific preferences who choose their strategies, interact, and produce an outcome.
In a complex adaptive system, however, the sequence does not end with the outcome. The effects of the interaction alter the environment in which the next interaction will take place. Actors react, learn, and adapt; their preferences may change, while new constraints and opportunities emerge and the available strategies shift. The outcome of an interaction thus becomes one of the initial conditions for the next interaction.
The conflict between Iran and the United States offers a significant example. U.S. military pressure has elicited an Iranian response that has progressively assigned the Strait of Hormuz a different centrality than it had at the outset of the confrontation. The restriction of traffic through the Strait has shifted the effects of the conflict from the military sphere to the energy and economic spheres, involving Gulf states, importing countries, markets, and shipping companies. These actors, in turn, have reacted by altering their behaviors, dependencies, and strategies. A decision made within the context of the military confrontation has thus helped transform the strategic environment in which Washington and Tehran must make subsequent decisions. Hormuz is no longer merely a geographical component of the confrontation, but a systemic lever capable of linking the military, energy, economic, political, and international relations dimensions.
The difference is substantial: the second move does not necessarily take place within the same game in which the first was made.
Actors learn, and alliances react. The conflict between Iran and the United States itself is helping to shift the regional balance of power. Saudi Arabia, while remaining committed to its strategic relationship with Washington, has strengthened security cooperation with Pakistan and intensified it with Turkey, thereby expanding and diversifying its strategic options. Actors who were initially outside the conflict are thus altering their behavior in response to the effects of the conflict and, in doing so, are helping to transform the environment in which Iran and the United States will have to make their next decisions.
Markets adjust prices, and populations change their attitudes. Technologies evolve, resources are consumed, and at the same time previously unknown vulnerabilities emerge, while actors who were initially peripheral gain prominence and reshape the range of available alternatives.
Strategic interaction thus becomes part of the process through which the system itself transforms. It not only produces consequences within a given environment but also helps reshape the environment in which the actors will continue to operate.
The system within which a state must make its next decision is therefore, at least in part, the product of previous decisions, including its own.
Adaptive Rationality
The first consequence concerns the concept of rationality. Game theory has gradually moved beyond the simple assumption of perfectly rational actors operating under conditions of complete information. Bayesian games allow for the representation of situations in which actors have incomplete information and form probabilistic beliefs about the characteristics, intentions, or strategies of others; evolutionary games study how different strategies can spread, compete, or disappear through processes of selection and adaptation; stochastic games introduce a dynamic dimension in which decisions and probabilistic events can alter the state of the game and thus the conditions of subsequent interactions. Incomplete information and bounded rationality have further expanded the scope of analysis, making it possible to model agents who make decisions without possessing all the information or without being able to process it perfectly. The problem posed by complexity, however, is yet another matter.
Under conditions of radical uncertainty, an actor may not know all possible future configurations of the system. Therefore, the actor must not only assign different probabilities to already identifiable outcomes but also consider that possibilities not included in the initial model may emerge and that interactions among actors may contribute to generating them.
This distinction is fundamental. In probabilistic uncertainty, we do not know which of the possible outcomes will occur; in radical uncertainty, we may not know all possible outcomes in advance. In a complex adaptive system, an additional element comes into play: the decisions of the actors help to modify the space of possibilities within which subsequent decisions must be made. W. Brian Arthur identified a similar problem while studying economic rationality in complex systems. When actors cannot deduce an optimal solution based on a complete and stable model of the environment, they construct partial representations of reality, formulate hypotheses, observe the effects of their decisions, learn, and progressively modify their behavior. Rationality thus takes on an adaptive character. The strategist cannot limit himself to seeking the optimal response to the current configuration. He must also assess how sustainable his choice will remain if that configuration changes and, above all, what possibilities for adaptation the current decision will retain for the future. The value of the outcomes (payoffs) may change
A second consequence concerns interests
In the simplest models of game theory, payoffs are defined before the game begins: they represent the value that each actor assigns to the various possible outcomes. In real-world international politics, however, interaction can alter not only the strategies through which actors seek to achieve their objectives, but also the relative value assigned to different outcomes. The war between Iran and the United States offers a significant example of this. What might initially have been conceived as a relatively brief confrontation has gradually turned into a war of attrition in which the energy, economic, and political dimensions have taken on increasing importance. The Strait of Hormuz, energy markets, the economic costs of the war, and the ability to sustain the conflict over time have thus become increasingly important components of strategic calculations.
Consequently, the outcomes that actors consider acceptable, advantageous, or too costly may not have the same value today as they did at the start of the conflict. It is not only the strategies used to achieve a given payoff that change; the value of the payoff itself may change.
The Relevant Actors are Also Changing
This transformation does not concern interests alone.
A seemingly bilateral conflict has consequences for actors who do not directly participate in the conflict. These actors react to the effects generated by the competition, and their reactions, in turn, alter the options available to the initial protagonists. Actors who were initially external or peripheral to the “ ” game can thus gain increasing relevance to the point of altering its configuration. The Strait of Hormuz is an almost paradigmatic example of this. Restrictions on shipping traffic do not affect only Iran and the United States but also involve Gulf exporters, Asian importers, European markets, shipping companies, and logistics chains. These actors do not merely suffer the consequences of the conflict—they adapt. Saudi Arabia, the United Arab Emirates, and other countries thus have further incentives to accelerate investments in oil pipelines, ports, storage capacity, and infrastructure capable of reducing their dependence on the Strait over time. The strategic consequences can be profound. Iran uses the Strait of Hormuz to increase its leverage in the present. The other actors respond by seeking alternatives. The development of different infrastructure, routes, and relationships can gradually reduce dependence on the Strait and thereby alter the strategic value of Iran’s leverage. The exercise of power can thus change the very conditions from which that power derives. This is one of the most significant paradoxes of strategy in complex adaptive systems: a capability used to gain an advantage in the present can trigger adaptations that reduce its effectiveness in the future.
From Feedback to Transformation
The process can be observed by following the sequence of effects. Military pressure on Iran contributes to Iran’s response regarding Hormuz; the reduction in energy flows increases costs; markets and states react; alternative routes and infrastructure are sought; dependencies gradually change, and through these, the very strategic value of Hormuz may change. Causality thus takes on a nonlinear and recursive structure. The effects produced by an action can in turn become causes of subsequent effects that feed back into the conditions from which the process originated. The literature on complex systems describes these dynamics through concepts such as feedback, nonlinearity, adaptation, and emergence. Interactions among agents can thus produce systemic effects that are not fully comprehensible by observing the behavior of each actor in isolation.
And this is where we must return, by analogy, to the metaphor of chess. In chess, a rook remains a rook regardless of how many times it is used. In international competition, however, the use of a capability can help alter its future value. A sanction can incentivize diversification. A blockade can accelerate the search for new trade routes. Technological superiority stimulates the development of countermeasures. An alliance can encourage the formation of counter-alliances. A dependency creates incentives to reduce it. Even the repeated use of deterrence can alter its credibility. The strategic system transforms itself through the actions, learning, and adaptations of the actors that comprise it.
The Problem of Equilibrium
The concept of equilibrium also requires caution. Equilibrium is one of the fundamental concepts of game theory: it allows us to identify configurations in which, given certain strategies, preferences, and information, no actor has an incentive to unilaterally change their behavior. In complex adaptive systems, however, it becomes equally important to understand the trajectory through which the system evolves, because the conditions under which an equilibrium is identified may themselves be subject to change. A relatively stable configuration observed today may therefore not represent an endpoint, but rather a temporary phase within an adaptation process. The Iran-U.S. conflict once again provides an example of this. Six months after the start of the war, neither side appears capable of fully imposing the desired outcome. Iran has been heavily hit but retains its resilience; Washington possesses incomparably greater military superiority but has failed to translate it into a definitive political outcome. Meanwhile, the confrontation has gradually expanded from the military sphere to sanctions, energy, maritime trade, and indirect negotiations. Rather than a balance in the technical sense, we are therefore facing a temporarily stable configuration of a system that continues to transform. And it is precisely this transformation that can progressively alter interests, the value of outcomes, available strategies, and the relevance of other actors. A balance captures a snapshot of a configuration; complexity compels us to observe its trajectory as well.
The Time Variable
This introduces a fundamental dimension: time. In repeated games, the future influences present behavior. If I expect to encounter the other party again, the benefits of cooperation, the value of reputation, and the incentives for reciprocity change. In complex adaptive systems, however, time plays an additional role.
It does not simply prolong the game: it allows actors to learn and adapt and, through their interactions, enables the system to transform itself. As the conflict drags on, arsenals are depleted and replenished, economies adapt, public opinion reacts, infrastructure is reconfigured, new technologies enter the fray, and alliances modify their behavior. The actors themselves learn from experience and can revise their objectives, priorities, and tools. The same strategy can therefore produce different results at different times because the system on which it exerts its effects is no longer identical.
This renders an assessment of power based solely on the resources possessed at the start of the interaction insufficient. Power must also be assessed through the ability to withstand the test of time, learn more quickly than others, adapt, and preserve options.
From the Optimal Move to Freedom of Action
This is where game theory and adaptive realism can converge. Realism rightly continues to remind us that states operate within a competitive system, that power matters, that security remains fundamental, and that the intentions of others can never be known with certainty. But in a complex system, power cannot be assessed solely by the quantity of available resources or the relative advantage held at present. What also matters is the ability to transform those resources into tangible effects without unduly compromising the ability to adapt to the resulting consequences.
The traditional strategic question—namely, what is the best response to the adversary’s move—must therefore be accompanied by a second question: which decision improves my current position without unduly reducing the options I might need when conditions change? This means balancing the pursuit of optimization with the preservation of future freedom of action.It does not mean giving up an advantage, nor does it mean replacing strategy with generic caution. It means recognizing that, in complex conditions, even the ability to continue pursuing one’s interests as conditions change constitutes a form of power. Diversification, redundancy, modularity, experimentation, feedback, reversibility, and portfolios of options are not, therefore, merely organizational criteria. They become components of strategy because they allow us to pursue current objectives without unduly constraining future possibilities.
A Strategy That Does Not Depend on Forecasting
The implication is significant. If the system is continually transformed by the interactions of its actors, no model can fully predict its evolution. This does not mean abandoning forecasting, but rather recognizing its limits and, above all, avoiding the confusion between forecasting and strategy. A robust strategy should not depend on the necessity of a particular forecast coming true.
Strategy in the face of uncertainty does not consist of perfectly predicting the future configuration. It consists of not building a strategy that requires the future to match our forecast in order to function. In the first case, we seek to identify the most likely future configuration and optimize our strategy accordingly. In the second, we build a position capable of pursuing our interests across a variety of possible configurations. This is the shift from optimization to adaptation. It does not mean preparing indiscriminately for any future—which is impossible and strategically inefficient. It means avoiding excessive dependence on a single hypothesis about the future and retaining sufficient options to correct one’s course when information, actors, constraints, or opportunities emerge that were not anticipated.
Beyond the Game
Game theory therefore remains essential. It has taught us that no strategic decision can be understood without considering the possible responses of others. Complexity theory, however, adds another layer. We must not only ask ourselves how the other party will react to our move. We must ask ourselves how the interaction between our move, their response, and the reactions of other actors will transform the system within which we will have to make our next decision. Interactions generate feedback; actors react, learn, and adapt; these adaptations alter behaviors, relationships, and constraints, thereby helping to transform the conditions of subsequent interactions. The game continues, but not necessarily under the same conditions under which it began. Chess remains an extraordinary representation of strategic intelligence. But contemporary international competition takes us further. In chess, we do not know our opponent’s next move, but we know the board, the pieces, and the rules. In strategic reality, we must consider that the second move may take place on a different board from the one on which we made the first.
Adaptive realism stems precisely from this awareness. Power continues to matter. Interests continue to matter. Competi continues to matter. What changes is the way we must evaluate them over time. Strategic advantage, therefore, does not belong solely to those who possess more resources or can calculate more moves ahead. It also belongs to those who can understand more quickly how the game is changing, adapt before others do, and retain sufficient options to continue pursuing their interests when the chessboard is no longer the one on which the game began.
This was first published in European Affairs and is published with the author’s permission.