Advancing Strategic Decision Science Since 2014
Human and economic interactions do not occur in a vacuum; they are embedded in networks of relationships—social networks, communication networks, trade networks, infrastructure networks. The Nevada Institute of Game Theory's research in network games investigates how this underlying structure of connections fundamentally shapes strategic behavior and outcomes. A network game is defined by a graph where nodes represent players and edges represent opportunities for interaction or influence. A player's payoff depends not only on their own action but also on the actions of their neighbors in the network. This framework has become indispensable for understanding phenomena as diverse as the adoption of new technologies, the spread of social norms, the dynamics of financial contagion, and the formation of alliances.
A key distinction in network games is between games of strategic complements and strategic substitutes. In complements, a player's incentive to take an action (like adopting a technology, working hard, or contributing to a public good) increases when more of their neighbors take that action. This leads to cascades and tipping points. In substitutes, a player's incentive decreases when more neighbors take the action (e.g., gathering information from friends—if many friends are informed, you can free-ride). NIGT researchers develop mathematical models to predict equilibrium outcomes in such games. They study how the network's topology—its density, degree distribution, presence of hubs, and community structure—affects the overall level of activity, the speed of diffusion, and the potential for multiple equilibria (e.g., a high-activity equilibrium and a low-activity equilibrium).
Network games provide a formal basis for analyzing peer effects. For example, in educational settings, a student's effort may depend on the efforts of their friends. NIGT models can help identify causal peer effects from observational data and design classroom or team assignments to maximize positive spillovers. Similarly, models of social learning on networks show how individuals update their beliefs based on the opinions of neighbors, potentially leading to informational cascades where society converges on a belief that may be wrong. In epidemiology, network models (though not strictly game-theoretic unless behavior is endogenous) show how disease spreads, but when individuals strategically decide to vaccinate or isolate based on the infection status of neighbors, it becomes a game. NIGT studies these 'behavioral epidemiology' models to design more effective public health interventions.
Perhaps the most fascinating aspect is that networks themselves are often the product of strategic decisions. Who do you befriend, trade with, or form an alliance with? NIGT researchers study network formation games, where players form or sever links based on the benefits (access to resources, information, support) and costs (maintaining a relationship) those links provide. Classic models include Jackson and Wolinsky's connections model and Bala and Goyal's two-way flow model. Analysis reveals which network architectures—like the star, the complete graph, or the core-periphery structure—are likely to emerge as stable or efficient. This helps explain the structure of business alliances, co-authorship networks, and international trade agreements. It also allows for the study of interventions, like subsidies for bridge-forming relationships, to improve overall network efficiency and resilience.
The applications are vast. In economics, network games model competition between firms located in a product characteristic space or interconnected through supply chains. In sociology, they formalize theories of social capital and influence. In computer science, they underpin the analysis of routing in the internet, the design of peer-to-peer protocols, and the study of viral marketing. The Nevada Institute's work in this area is highly interdisciplinary, drawing on graph theory, sociology, economics, and physics. By rigorously modeling how structure begets strategy and strategy begets structure, network game theory provides a dynamic and holistic view of our interconnected world, allowing us to predict, and perhaps one day design, social and economic networks that foster cooperation, innovation, and collective well-being.