Methodology

This is a Monte Carlo simulation — a technique that runs thousands of randomized scenarios through a mathematical model and shows you the distribution of outcomes. Instead of predicting a single future, it shows you the range of possibilities and how likely each one is.

The model behind this simulation is a Bayesian network: a directed graph of 107 interconnected variables spanning 8 phases of a potential US-Iran military conflict. Each variable has a probability distribution that depends on its parent variables. When you change an input — like selecting a crisis trigger scenario — the effects cascade through every connected variable in the network.

Each variable in the network (called a “node”) has a conditional probability table (CPT) that defines how its probability distribution changes based on the state of its parent nodes. For example: if Iran's nuclear program is at breakout threshold AND IRGC military readiness is high, the probability of an escalation trigger event rises from a baseline of 15% to approximately 45%.

When the CPT doesn't have an exact match for the current combination of parent states (which happens frequently with sparse data), the engine uses weighted interpolation — blending the closest matching CPT entries proportional to how many parent values they match. This ensures that parent-child relationships always influence outcomes, even when the data is incomplete.

Each simulation run samples values for every node in topological order (parents before children), building one complete scenario from tensions through aftermath. The default configuration runs 5,000 independent scenarios to produce the probability distributions you see in the results panel.

This simulation's conflict dynamics are informed by Robert Pape's empirical research on strategic coercion, the most comprehensive scholarly analysis of airpower's political effectiveness. Pape analyzed every strategic air campaign from 1917 to 1991 (33 cases) and found a striking pattern: airpower almost always succeeds tactically but almost never achieves its political objectives.

The denial/punishment distinction.Pape classified coercive airpower into four strategies. Punishment (targeting civilians to force capitulation) failed in all 8 cases where it was the dominant approach. Decapitation (killing leaders) has never worked independently against states. Only denial (threatening the enemy's ability to achieve military objectives) produced success — and only 15% of the time (5 of 33 cases), always requiring ground forces to create what Pape calls a “hammer and anvil” effect.

The escalation trap.Applied to Iran, Pape's framework predicts a three-stage escalation pattern. Stage 1: precision strikes achieve tactical success but fail to eliminate dispersed nuclear knowledge. Stage 2: strategic disappointment drives escalation to regime-change bombing. Stage 3: air coercion fails to produce political compliance, creating structural pressure for ground forces (~75% probability by Pape's estimate). The core logic is recursive: failure generates fear, fear justifies escalation, escalation produces new failure.

Nationalist backlash.Pape's most counterintuitive finding: bombing consistently strengthens defender resolve rather than degrading it. This is the inverse of what punishment theory predicts. Authoritarian regimes suppress internal revolt under external pressure, and civilian casualties trigger rally-around-the-flag effects. Iran absorbed hundreds of thousands of casualties during the Iran-Iraq War without collapsing. The simulation models this as a positive feedback loop where higher bombing intensity increases Iranian nationalist resolve, making political concessions less likely.

How this shapes the model.Four nodes directly encode Pape's parameters: escalation trap stage (Phase 3), Iranian nationalist resolve (Phase 6), coercive success probability (Phase 7), and ground escalation probability (Phase 7). Resolution outcomes are calibrated against Pape's empirical success rates — “decisive US victory” through airpower alone is set near 5%, consistent with the 0-15% historical range for punishment and denial strategies.

Limitations of the framework.Pape's 33-case dataset predates precision-guided standoff weapons, advanced cyber capabilities, and Iranian drone swarms. His bilateral model does not fully capture multilateral dynamics (US, Israel, Iran, proxies, Russia, China). The ~75% ground escalation estimate is a single analyst projection without a historical base rate for this specific scenario. Competing scholars (Kroenig, Mueller) argue Pape underestimates punishment's marginal contribution in compound strategies.

Each result card shows a confidence badge — high, medium, or low — computed from three factors:

• CPT Coverage (40%): How often the simulation engine found a matching or partially-matching CPT entry versus falling back to default probabilities.
• Sample Adequacy (30%): Whether there are enough simulation runs to produce statistically meaningful results for this variable.
• Source Quality (30%): The quality of the underlying data sources used to calibrate this variable.

High confidence: Solid data, good model coverage, and sufficient samples. Trust the distribution shape and approximate values.

Medium confidence: Directional estimate — trust the trend but not the exact percentages.

Low confidence: Speculative — the model is making educated guesses. Treat as hypothesis generation, not forecasting.