Run downside case
Explore what happens if the important assumptions move against the forecast.
Scenario / Simulation LabSynthetic evidence only
Explore possible futures without turning simulations into truth.
Scenario Lab is the User-facing window into KELLY's Monte Carlo and uncertainty work. You can request bounded simulations, compare assumptions, and inspect receipts. You cannot steer KELLY's belief system with synthetic output.
Compare scenarios
Place a base, optimistic, and stress case side by side without hiding the assumptions.
Show confidence interval
View uncertainty bands only when the method declares the kind of uncertainty being shown.
Stress this forecast
Ask what would need to change before a forecast becomes fragile or no longer useful.
Correct path
Simulation can inform. Reality must settle.
01
Scope
KELLY checks whether the request is bounded and safe.
02
Simulate
The Monte Carlo organ can produce synthetic scenario evidence after admission.
03
Receipt
A run gets hashes, provenance, labels, and replay class.
04
Inspect
Glass Lab shows the run and why it is not empirical proof.
05
Settle later
Only reality closure and calibration can upgrade future memory eligibility.
Current state
This is a research-preview shell. MCE remains protected: users can request bounded scenarios and inspect the proof rules, but raw engine controls, authority mutation, and memory promotion stay gated behind KELLY-side admission.