A Full Conflict Simulation in 395KB: CENTCOM War Game Runs Entirely in Your Browser
What happens when you point an AI coding assistant at declassified military data and ask it to build a wargame?
You get CENTCOM War Game β a full-spectrum conflict simulation that runs entirely in your browser. One HTML file. 395KB. No backend.
What It Does
This isnβt a game in the entertainment sense. Itβs an analytical tool that models a hypothetical US/Iran conflict using real-world data:
Data Sources (OSINT):
- IISS (International Institute for Strategic Studies) β military ORBAT
- CSIS (Center for Strategic and International Studies) β strategic analysis
- EIA (Energy Information Administration) β oil market data
- IAEA β nuclear program estimates
- IMF β economic baselines
Modeling Approaches:
- Lanchester combat model β classic attrition warfare mathematics
- Game-theoretic escalation β decision trees for conflict escalation
- Monte Carlo analysis β 100 simulation runs for probabilistic outcomes
- Full-spectrum conflict β military, economic, cyber, proxy warfare
The Scenario Presets
The simulation comes with five preset scenarios:
| Preset | Description |
|---|---|
| Pre-War | Full Iranian military intact, diplomatic tensions |
| Epic Fury D1 | Opening strikes, leadership decapitation |
| Epic Fury D5 | Navy/Air Force destroyed, IADS neutralized |
| Epic Fury D7 | Full air superiority, mop-up phase |
| Ground Invasion | What-if: US escalates to ground war |
Each preset adjusts force dispositions, economic conditions, and escalation parameters. Run 100 Monte Carlo simulations and watch the probability distributions shift.
The Technical Implementation
Whatβs impressive isnβt the politics β itβs the engineering:
Total size: ~395KB
βββ Leaflet.js (interactive map)
βββ Chart.js (visualization)
βββ 6 OSINT data files (embedded as JS objects)
βββ Lanchester combat engine
βββ Monte Carlo simulation loop
βββ LiveUAMap-style UI
Everything runs client-side. No server calls except to load the page. You could save it locally and run simulations offline.
Calibrated parameters:
- Oil price range: $80-125/barrel
- Casualty ratios: 12-57:1 (Iran:US)
- Nuclear threshold: cumulative-to-daily rate conversion
Built with Perplexity Computer
The repo credits Perplexity Computer as the development tool. This matters.
Weβre entering an era where AI coding assistants can ingest research-grade data (military ORBATs, economic models, conflict theory) and produce functional analytical tools in hours rather than months.
The implications:
- Democratized policy analysis β Think tanks and academics no longer have monopoly on simulation tools
- Rapid scenario modeling β Spin up a new conflict simulation as events unfold
- Transparent methodology β Open source means the models can be audited and criticized
What This Isnβt
Letβs be clear: this is a simplified model. Real military planning uses classified data, proprietary simulations, and decades of institutional knowledge. A browser-based tool with OSINT data isnβt replacing RAND Corporation or the Pentagonβs wargaming division.
But itβs a proof of concept for something new: analytical tools that anyone can run, fork, and modify.
The Bigger Picture
Single-file applications are having a moment. Weβve seen:
- zclaw β 888KB AI assistant on ESP32
- SkyClaw β 6.9MB AI agent binary
- Now: 395KB conflict simulation
The pattern: pack maximum capability into minimum payload. No dependencies, no deployment complexity, no ongoing costs. Just download and run.
For analytical tools specifically, this matters because it removes barriers to access. A policy student in Lagos can run the same simulation as a defense analyst in Washington. The data is open. The models are transparent. The compute is your browser.