ExpKorzhCow
Tier B: Korzh et al. 2015, COW over 307 km of ultra-low-loss fibre
11/11 passed in 1ms
| Test | What it does | Result |
|---|---|---|
test_korzh_ceiling | Their secret rates sit at 12.1% (307 km) to 45.0% (203 km) of the per-click ceiling; declared miss: against their sifted block the ceiling reads 0.672 and 0.689 at 203 and 307 km. | ✅ pass |
test_korzh_dark | Background-free the model is flat at 1.1537%; halving their dark column puts 307 km at 2.36% and doubling it at 5.63%, both excluded by their 3.5%. | ✅ pass |
test_korzh_dead | Model click rate plus measured dark counts uses 61.9% of the 1/dead_time ceiling at 104 km, falling to 0.32% at 307 km. | ✅ pass |
test_korzh_form | cow_rate is | ✅ pass |
test_korzh_near | Declared miss: 104 km reads 1.526% against their 2.40%, needing an afterpulse-like background of 1.84% of the count rate, against 0.63% at 256 km and -0.11% at 153 km. | ✅ pass |
test_korzh_penalty | Their Eq. (3) fraction less the 7 sqrt(n_cpp log2 1/beta) term reproduces f_sec to 0.05% at 104 km (block 1.97e7) and misses by 38% at 307 km (6.6e5), the unmodelled Eq. (2) visibility deviation. | ✅ pass |
test_korzh_qber | Fitted at 203 and 307 km, the model gives 1.555% at 153 km against their 1.5% and 1.700% at 256 km against their 2.0%. | ✅ pass |
test_korzh_receiver | The fitted factor over their 20% efficiency needs a 0.9331 data-line share, a lower bound since the 20% is stated below 150 K; refitting at 153 and 307 km moves the factor 2.8% and the baseline 5.8%. | ✅ pass |
test_korzh_secret | qkd's asymptotic COW fraction on the Branciard-Gisin-Scarani phase error exceeds their finite-key f_sec at every distance, 1.065 times at 104 km and 2.133 at 307 km. | ✅ pass |
test_korzh_span | The loss column exceeds bare 0.160 dB/km fibre by 0.26 dB at 104 km rising to 2.78 dB at 307 km, 0.1625 to 0.1691 dB/km end to end. | ✅ pass |
test_korzh_table | Source defect: r_sec = n_cpp f_sec / t_cpp holds to 1.2% at 104, 203 and 307 km but reads 0.810 at 153 km and 0.446 at 256 km. | ✅ pass |