skip to content
Scenario explorer

Expected future, beside the credible unexpected

Deterministic reference models and bounded perturbations. Nothing here is a trained forecast — it is an orchestration of synthetic references.

Demo modelSynthetic dataResearch preview

pipeline stage 05

Opened from How it works with the Financial Services episode preloaded and a sensor failure injected, so the blindness channel is separated from genuine calm. Advance the clock and watch the credible tails separate from the expected trajectory.

Scenario templates

Preloaded episodes — assumptions, clock position and injections set for you. Every control below stays editable.

Domain

Initial state

Undercollateralized credit line — synthetic borrower 0x…8fa2

Token wealth93.8 index
Stablecoin cash flow55.7 index
Leverage3.3 x
Counterparty concentration42.8 HHI

Assumptions

These change how consequence is read, not what the references predict.

exposure · h1 · near×1.00
exposure · h4 · medium×1.00
exposure · h10 · long×1.00
credibility floor0.15

Tails below the floor stay visible but are marked discounted — a tail is never silently deleted.

risk aversion0.50

0 reads consequence credibility-weighted; 1 reads it worst-case, ignoring how likely the tail is.

Time & injections

observation 6/12