USA vs Iran: Conflict Model

Game theory simulation of the US-Iran military conflict (Operation Epic Fury) with 4 players, stochastic shocks, oil dynamics, and nuclear escalation. Updated with 36 days of real-world data.
v6 · Day-36 warm start · 500 Monte Carlo simulations · Recalibrated from reality

How the Model Works

Each "round" represents ~2 weeks. Four players choose strategies simultaneously, like a poker game where everyone reveals their hand at once.
1

4 Players

USA (air strikes, ground ops, or withdraw), Israel (independent or joint), Iran (retaliate, blockade Hormuz, or endure), Hezbollah (rocket barrages or ceasefire)

2

Each Optimizes Payoffs

Every player picks the strategy that maximizes their expected benefit, considering military losses, domestic politics, oil prices, and nuclear risks

3

Random Shocks

Unexpected events: Iraqi militia attacks, Hormuz mine incidents, IRGC leadership kills, embassy seizures, tanker attacks — each changes the calculus

4

500x Monte Carlo

We run 500 simulations with different random seeds. The distribution of outcomes shows probabilities, not a single prediction

Base Prediction: Who Wins?

Distribution across 500 simulations with a revolutionary (risk-taking) Iran
Coalition victory
Iran strategic victory
Negotiated / Frozen

Model vs Reality Day 36

After 36 days of war, how did the model's predictions compare to what actually happened?

Model Got Right

Model Got Wrong

Recalibrations Applied (v6)

    From Here: Day 36 Forward Projection NEW

    Starting from the actual war state on April 5, 2026 (Day 36): Iran military ~28%, Hormuz blocked, oil $105, Trump rhetoric at 0.6. What happens next? (500 simulations)
    Coalition victory
    Iran strategic victory
    Negotiated / Frozen

    Day 36 What-If Scenarios

    How do different assumptions change the outcome from the current war state?
    Scenario Coalition % Iran % Avg Days Oil $/bbl Nuclear %

    How Wars End: 6 Archetypes

    Not all coalition victories are equal. The model identifies distinct "war stories"

    Simulated Wars, Step by Step

    Each button shows a different simulation run. Same model, different random seed — different war. Compare how small changes in luck lead to completely different outcomes.

    Key metrics over time

    What If Trump Is More Patient?

    The "discount factor" measures how much a player values future gains vs. present costs. Lower = more impatient, wants quick results
    Impatient (wants quick win) 0.82 Patient (like Bush post-9/11)

    What If Oil Is Already Expensive?

    Starting oil price changes everything: higher oil = Iran's Hormuz blockade becomes more painful for the US
    Low oil ($75) $85 High oil ($130)

    Scenario Comparison

    Side-by-side comparison of all what-if scenarios
    Scenario Outcome Distribution Coalition % Iran % Avg Rounds Avg Oil $ Nuclear % Recession

    Key Findings

    What does the model tell us?
    1. Coalition wins most scenarios (~73%) but the most common outcome (46%) is a "Pyrrhic victory" with oil >$150 and significant economic damage.
    2. Oil price is the single biggest lever. Starting at $130 instead of $85 flips the odds: Iran wins 58% vs 27%. Iran's strategy of blocking the Strait of Hormuz becomes devastating when oil is already expensive.
    3. Trump's patience matters a lot. An impatient Trump (quick withdrawal) gives Iran a ~33% chance. A patient "Bush-like" commitment drops it to ~20%.
    4. Nuclear breakout paradoxically helps the coalition. When Iran approaches nuclear capability, it triggers escalation that accelerates Iran's military destruction. But it creates long-term instability.
    5. Short wars favor Iran. If the war ends in <8 rounds, Iran wins 100% of the time — because the only reason for a short war is US withdrawal.
    6. "Managed degradation" (clean win) happens only 19% of the time. In most scenarios, even winning is expensive.

    Want to Dig Deeper?

    The full model is open-source. You can run it yourself, change parameters, or build on it.

    Read the Code

    The core model is a single Python file (~600 lines). It implements a 4-player repeated game with Nash equilibrium selection, stochastic shocks, and oil price dynamics. No ML — just classical game theory.

    game_theory_conflict_model.py →

    Run Your Own Scenarios

    Clone the repo, change parameters (patience, oil price, shock probabilities, military capabilities), and run 500 simulations on your laptop in under a minute.

    github.com/mool32/game_theory_models →

    Key Assumptions

    Each player maximizes a discounted payoff function that weighs military gains, casualties, oil costs, domestic support, and nuclear risks. Iran's "revolutionary" archetype overweights regime survival. Strategies are selected via best-response dynamics (approximate Nash).

    What the Model Doesn't Capture

    Diplomacy, intelligence operations, cyber warfare, internal regime politics, Chinese/Russian involvement beyond oil price effects, humanitarian costs, long-term occupation dynamics. It's a simplification — useful for intuition, not prediction.

    Disclaimer

    This is an exploratory game-theoretic model, not a forecast. Real conflicts involve thousands of variables that no model can capture. The probabilities shown reflect the model's internal logic, not objective likelihoods. The model is useful for understanding dynamics and sensitivity, not for prediction.