PairModel
Photon-pair gain and error rate against an independent enumeration
8/8 passed in 2ms
| Test | What it does | Result |
|---|---|---|
test_asymmetric | Gain and error rate agree with the enumeration at three unequal arm pairs. | ✅ pass |
test_background | Ma-Fung-Lo carry no background factor on the correlated term, so the QBER is exact at Y0 = 0 and rises with the background. | ✅ pass |
test_brightness | The optimal mean pair number falls over 25 dB but stays within half its back-to-back value, which is Ma-Fung-Lo's stated result. | ✅ pass |
test_correlated | The correlated fraction is the closed form 2*(1/c - 1/(a*b)) and never exceeds the gain. | ✅ pass |
test_exact | Gain and error rate reproduce a four-detector enumeration at zero background, which fixes Ma-Fung-Lo Eq. (10)'s denominator exponent -- the squared reading misses by 7% at lam = 0.1. | ✅ pass |
test_forecast | The forward CHSH model reaches Ekert's Eq. (4) at his own settings, its correlations, S and visibility bridge being Eqs. (2), (3) and pr.chsh. | ✅ pass |
test_limits | A vanishing brightness leaves the misalignment as the error rate, and a link with no coincidences reads the chance rate 1/2. | ✅ pass |
test_scaling | The optimised rate is linear in the end-to-end transmittance, log-log slope 1.0 within 0.05 over 6 to 25 dB. | ✅ pass |