Skip to content

Scelo Atlas

Scelo Atlas — synthetic risk lab, Karl's world, priced live

World-scale risk simulation for actuaries — an Unreal Engine 5 plugin by Intelligent Actuaries.

What Twinmotion is to architects, Scelo Atlas is to actuaries.

One draw from 40,000 years — the EP curve over Karl's world

Scelo Atlas brings real catastrophe modeling into Unreal Engine 5: parametric hazard physics (Holland wind fields, tornado vortices, flood inundation, ground-motion attenuation), insured exposure portfolios, engineering vulnerability curves, a background-threaded Monte Carlo event-loss engine producing EP curves / AAL / TVaR — and a cinematic layer where a hurricane visibly tracks across a portfolio while gross losses accrue in real time on screen.

It is the simulation front-end of Scelo, the actuarial intelligence platform by Intelligent Actuaries. The optional Scelo Bridge ships each simulation run to Scelo's climate and capital agents and returns an AI-written executive brief inside the engine. Without the bridge the plugin is fully offline; every result exports to local JSON and CSV.

The invariant that matters

The cinematic demo and the 100,000-year batch run share one hazard model and one financial engine. Live events accrue dollars as a running max of hazard intensity over exactly the analytic swath's discrete samples, so end-of-event live totals equal the analytic event totals bitwise — logged as a floating-point-exact assert. What the executive watches is what the actuary computed.

A live tornado repricing the book in the Scelo Lab

Why this matters

Actuaries today see catastrophe risk as rows in a spreadsheet. Scelo Atlas lets a Chief Risk Officer stand inside the 1-in-250 event: watch the storm make landfall over their actual portfolio, watch the loss counter cross the reinsurance attachment point, then read the EP curve the engine computed from 100,000 simulated years. Same rigorous math — radically better intuition, communication and decision-making.

What's in v0.2

Capability Detail
Hazard physics Holland (1980) asymmetric cyclone wind field, Rankine tornado vortex, parametric flood inundation, tsunami surge, point-source PGA attenuation, playback of real strong-motion earthquake records
Exposure CSV portfolio import (lat/lon, elevation, TIV, occupancy, policy terms), OED import, geo-projection, instanced-mesh portfolio visualizer
Vulnerability Per-(peril, occupancy) curves with Beta-distributed secondary uncertainty
Financial engine Site deductibles/limits/shares, event-loss tables with ground-up vs insured split, OEP & AEP curves, AAL, annual σ, TVaR, reinsurance and parametric structures, capital allocation
Monte Carlo Poisson (or negative-binomial) event frequency over a stochastic catalog, background-threaded, 100k+ years, progress + completion delegates
Scelo Lab Sandbox arena where every shape you build is priced as an insured location — see Scelo Lab & Climate Studio
Cinematic layer Live cyclone/flood/tornado/quake actors plus a one-command ~66 s film — see Cinematic
Climate Studio Pathway × horizon scenario tool with adaptation levers (illustrative)
Exports Full actuarial results pack: ELT, YLT, PLT, EPT, locations, footprints, metrics JSON — see Bridge & Exports

Documentation

  1. Installation
  2. Quick start
  3. Concepts & architecture
  4. Data formats
  5. Console reference — all 90+ Scelo.* commands
  6. Cinematic
  7. Scelo Lab & Climate Studio
  8. Scelo Bridge & exports

Requirements

  • Unreal Engine 5.8 (compiles and runs clean on 5.8.0, Windows/MSVC; module allow-lists also cover Mac and Linux)
  • C++ project (the plugin compiles once on first build)
  • Engine plugin dependencies Niagara and Water are declared and enable automatically

© Intelligent Actuaries. All rights reserved. · intelligentactuaries.com