ExpHajomerRelay

Tier B: Hajomer et al. 2025, CV-MDI measured over 10 km of fibre

5/5 passed in 0ms

Test	What it does	Result
test_cvmdi_floor	The pure-loss equivalent noise at ${\tau }_A=1$, ${\tau }_B=0.56$ is 5.5714, against 4.3457 for the most favourable correlated environment and their measured 5.6815.	✅ pass
test_cvmdi_phase	Their phase-noise form is form="literature" term for term, 12.307 mSNU at 0.06 rad against their 12.6, inside PHASE_LIMIT with no PhaseDomainWarning.	✅ pass
test_cvmdi_rate	Their 39.5 mSNU referred to Bob's arm through ${\tau }_B$ gives a 0.180 bit/symbol per-attack rate at their $\beta$, above the 0.130 their 2.6 Mbit/s at 20 MBaud comes to.	✅ pass
test_cvmdi_referral	Every referral of their relay-plane 39.5 mSNU, from all on Bob to an even split, and their worst-case estimator, stays above their 0.130 bit/symbol.	✅ pass
test_cvmdi_relay	Charging the relay's 94% efficiency to Eve as arm loss, instead of calibrating it out as their ${\tau }_A=1$ does, takes the rate to -0.134 bit/symbol.	✅ pass