Numerical experiment series

Six studies on a cat state.

Each study below is computed in your browser when this page loads — a fresh run against the same state the chamber holds. Nothing here is a stored table. Change the seed and every number changes; keep the seed and the run reproduces exactly, down to the last decimal.

What this is. These are numerical experiments: real mathematics executed on a simulated quantum state, in the sense that computational physics is experimental. They are not laboratory measurements — no ion was trapped, nothing was cooled, no photon was counted. Where a result is compared against theory, the theory is cited and the comparison is made honestly, including when the agreement is only approximate.
Run
Seed
State
Grid
Wall time
H-01

Hardware run

Loading…
Superconducting QPU · IBM Quantum · GHZ populations + parity oscillation

Method

The state under test

An equal superposition of N coherent states on a ring in phase space, αk = α·ei(2πk/N+φ), optionally squeezed and damped. Its Wigner function is evaluated in closed form on a 96×96 grid with all cross terms retained — those cross terms are the interference, and dropping them would turn the state into an ordinary mixture.

Reproducibility

All stochastic parts run on a seeded mulberry32 generator, never Math.random. The seed appears in the URL, so a run is a link. Re-running the same seed on the same build reproduces every digit; the exports carry the seed, the state parameters and the wall time alongside the results.

What would falsify it

Each study declares a prediction from published theory and a pass condition before it runs. A KS p-value that collapses, a Wehrl excess below Lieb's bound, a Mandel Q that goes negative for a coherent state, or a fringe exponent away from 1 would each mark the run failed — and the badge says so rather than hiding it.