Evolvix Prototype Compiler — Download and RiskyMAD Model Code#
The tool behind the nuclear winter forecast. Download it. Run the model. Check the math.
Download#
The Evolvix prototype command-line compiler (MMv0r3p1-RC1, 2015m03d11) is the tool used to produce the stochastic RiskyMAD simulations in SD1.
Platform |
Size |
File |
|---|---|---|
macOS (OS X 10, x64) |
6.2 MB |
|
Windows 7 (x64) |
5.9 MB |
|
Ubuntu 14 (x64) |
14.3 MB |
|
Fedora 21 (x64) |
14.5 MB |
|
RHEL 7 (x64) |
14.3 MB |
|
User documentation (HTML) |
3.4 MB |
|
Intro manual (PDF) |
739 KB |
evolvix-prototype-intro-manual-mmv0r3p0beta_2014m10d01-57page.pdf |
License: Jonah License with CC0 Public Domain
Evolvix project archive: web.archive.org/web/20191227220523/http://evolvix.org/
For the full Evolvix vision and research roadmap, see STa1-EVX.
The RiskyMAD Model Code#
The complete Evolvix model code as shown on the SD1 poster. This is the point estimate used for the base-case forecast (median ~19 years to accidental nuclear winter):
Evolvix Quest RiskyMADdead
(Question: "How many years until humanity self-destructs
in a nuclear roulette accident?")
!L works with prototype Evolvix MMv0r3p1-rc1
Simulate stochastically until 200 :["years"]
!L Run 1 random world history into future
!L Simulate deterministically until 200 :["years"]
!L Predict future of countless averages
Initial Amount of Risky = 1 !L starts as 1 = active now
Initial Amount of MAD = 0 !L only 1 if reached from Risky
Initial Amount of Dead = 0 !L starts at 0, becomes 1 in nuclear winter
Initial Amount of rRiskyGoMAD = 0.10 !L Global rate: 0.03-0.3 /year
Initial Amount of rMADescapes = 6 !L Oversimplified or overcomplicated escape
Initial Amount of rMADtoDEATH = 3 !L Non-escape: rate of overreach
Action 1 ( Risky ---[ Rate = 0.10 ]---> MAD ) !L crisis arises
Action 2 ( MAD ---[ Rate = 6 ]---> Risky ) !L crisis de-escalates
Action 3 ( MAD ---[ Rate = 3 ]---> Dead ) !L nuclear exchange
Action 4 ( Risky ---[ Rate = 0 ]---> LifeMAP ) !L escape (currently 0)
How to read this code: Evolvix models stochastic systems using the Gillespie algorithm (exact stochastic simulation of continuous-time Markov chains). The model has three states (Risky, MAD, Dead) and four transitions (Actions 1–4). When the system is in the MAD state, two competing reactions fire stochastically: escape back to Risky (rate 6) or death (rate 3). The probability of death per crisis is therefore 3/(6+3) = 1/3. The escape rate (Action 4) is set to 0 — no MAP transition is currently active.
Why these rates? The crisis-resolution rates (6 and 3) are grounded in the BABL death-trifecta ([Matheo-2]): under BABL, a crisis resolves through one of three modes — over-Simplifying (returns to Risky), over-Complicating (returns to Risky), or over-Reaching (leads to Dead). Two out of three modes produce escape; one produces death. The factor of 3 sets the crisis time scale (~40 days average duration, consistent with historical crises).
Simplicity is the point. In other simulation frameworks, building this model would require hundreds of lines of code, custom data structures, and significant programming expertise. In Evolvix, the entire model fits on a poster. Anyone who can read the code can check the math. This is what Evolvix was designed for: making accurate stochastic modeling accessible to non-programmers. #AuditTheMath