Human-Machine Strategic Analysis: New Tools for Behavioural Models and Security Decision-Making is a two-part series exploring how network analysis, agent-based modelling, wargaming, and AI decision support can be combined with behavioural and political psychology to understand and responsibly design human-machine judgment in security and foreign policy. Serie A offers six accessible introductions to the core tools and debates. Series B is a set of eleven in-depth, technical essays, translating behavioural theory into modelling design.
Series A – Public-Facing Essays
Six Accessible Essays introducing the core tools and debates
Network Analysis: Seeing Power as Position
Networks · Power · International Relations
Power in international politics often comes from where an actor sits in a system of dependence, access, and flow, and not from resources counted in isolation. Chokepoints, sanctions regimes, proxy networks, and strategic
corridors show power as position, not stockpile.
Agent-Based Modelling: Strategy as Interaction, Not Linearity
ABM · Emergence · Strategic Interaction
Strategic interaction resists linear cause-effect explanation because actors adapt, learn, imitate, and retaliate.
ABM’s value lies not in prediction but in showing how local, rule-governed interactions between actors produce
macro-level outcomes nobody designed.
Wargaming: What Games Can and Cannot Tell us
Wargaming · Judgment · Simulation
Wargaming explores human judgment under uncertainty, not the future. Its value lies in revealing assumptions,
decision pressures, misperceptions, role conflicts, and escalation pathways, which depends entirely on
adjudication, scenario design, and player interpretation.
AI Decision Support: Useful Assistant or False Confidence Machine?
AI · Decision Support · Human–Machine
AI may help with synthesis, scenario exploration, red-teaming, and structured comparison, but strategic
judgment still requires human responsibility and awareness of uncertainty. The real question is how humanmachine teams should divide analytical labour.
Why Prediction Is the Wrong Standard for Strategic Models
Models · Epistemology · Decision-Making
Strategic models should not be judged only by predictive accuracy. Models are useful if they clarify
assumptions, reveal mechanisms, expose blind spots, generate scenarios, or structure disagreement.
Distinguishes prediction, explanation, exploration, and decision support as four separate jobs a model can do.
Human-Machine Judgment in Strategic Analysis
Human–Machine · Judgment · Strategic Analysis
Humans suffer bias, limited attention, motivated reasoning, and group dynamics; machines encode assumptions
and can mislead with false confidence. The task is designing environments where each compensates for the
other’s weaknesses.
Series B – Technical Model-Card Essays
Eleven in-depth Essays translating behavioural theory into modelling design
This series asks what artificial political actors must perceive, remember, value and do for behavioural theories to operate inside a simulation.
Essay 1 – Bounded Rationality and Political Agents
Bounded-rational agents operate with limited information, limited attention, and simplified decision rules rather than full optimisation. Relevant for modelling voters, leaders, protesters, bureaucracies, and armed groups.
Essay 2 – Heuristics and Biases as Agent Rules
Availability, representativeness, anchoring, and status-quo bias can become agent decision rules shaping how simulated actors perceive risk, choose targets, or imitate others. However, poorly specified biases risk becoming arbitrary “psychological decoration.”
Essay 3 – Prospect Theory and Risk in Crisis Decision-Making
Actors respond asymmetrically to gains and losses (risk-acceptant facing losses, risk-averse protecting gains) which may explain escalation under domestic pressure, territorial loss, regime insecurity, or perceived humiliation.
Essay 4 – Groupthink, Polythink, and Collective Decision-Making
Foreign-policy decisions emerge from advisory groups and bureaucratic bargaining, not unitary actors. Groupthink produces premature consensus; polythink produces fragmentation. ABM and wargaming can represent advisory structure, disagreement, and information bottlenecks.
Essay 5 – Emotions, Threat, and Political Mobilization
Fear, anger, humiliation, hope, and moral outrage shape participation, repression, threat perception, and escalation. Emotion can be represented as a mechanism altering attention, risk tolerance, group solidarity, and willingness to act.
Essay 6 – Identity, Norms, and Cultural Models in Simulation
Identity, norms, and culture shape perception, trust, cooperation, and threat, and should be modelled as relational and situational mechanisms, not fixed essences. The series’ highest-risk, most carefully sourced piece.
Essay 7 – Motivated Reasoning and Information Environments
Actors interpret new information through prior beliefs, identity, and incentives rather than updating neutrally. Relevant for propaganda, crisis misperception, intelligence failure, and protest diffusion.
Essay 8 – From Behavioural Theory to ABM Design
A translation pipeline for turning any behavioural theory into a modelling decision: identify the theory, define the actor, specify the decision environment, formalise the decision rule, define interaction mechanisms, run scenarios, compare emergent outcomes.
Essay 9 – Decision Rules and Poliheuristic Theory
Once bounded rationality has narrowed a decision to a manageable set of options, how does an agent actually choose among them? Noncompensatory decision rules and poliheuristic theory’s two-stage screen-then-select process give a formal, testable answer, with direct payoffs for how an ABM or AI decision-support agent should be programmed to choose.
Essay 10 – Framing and Counter-Framing
The same objective set of facts can be made to look like a gain or a loss, a free-speech issue or a public-safety issue. Which framing wins in a competitive elite environment shapes public opinion and elite choice independently of any change in the underlying facts. Directly relevant to human-machine strategic analysis: an AI decision-support system that presents options to a human decision-maker is, whether or not it intends to be, a choice architect and a framer.
Essay 11 – Nudging and Choice Architecture
Choice architecture, defined as the deliberate design of the context in which people choose, can shift behaviour at scale without eliminating options or changing incentives. Directly relevant to human-machine strategic analysis: any AI decision-support interface is itself a piece of choice architecture.
How the Two Series Connect
Representations of the system + representations of the actor + human interpretation = strategic analysis
The first series asks how analytical tools represent strategic environments. The second asks how models represent the political actors operating within them.
Neither level is neutral. Strategic analysis emerges from the interaction between representations of the system, representations of the actor, and the human judgment used to interpret both.
What should remain human? What can be delegated? And what happens when neither human nor machine should reason alone?