p=1e-3, n=50: independent 4.9e-02, union bound 5.0e-02
harmonic register 1, 2, 3: parity dimension 2, smallest angle 31.9 deg
drift calibrated, two colors + polarimeter: parity dimension 1
drift calibrated, + path sensor: parity dimension 2
drift free, two colors + polarimeter: parity dimension 0
drift free, + path sensor: parity dimension 1
drift free, + path sensor + drift prior: parity dimension 2
drift calibrated + path sensor, noise-weighted: smallest angle 7.4 deg
window of 1 samples, constant drift: parity dimension 0
window of 2 samples, constant drift: parity dimension 1
window of 3 samples, constant drift: parity dimension 2
window of 4 samples, constant drift: parity dimension 3
  10.59 um: N_F = 2.105e+20 m^-2
   5.22 um: N_F = 4.271e+20 m^-2
   4.60 um: N_F = 4.847e+20 m^-2
DIII-D prototype 10.59/5.22
  (1, 2): dN = +0.0187 fringe (+3.94e+18 m^-2), dL = +10.392 um
  (3, 7): dN = -0.5950 fringe (-1.25e+20 m^-2), dL = +38.071 um
  (10, 23): dN = -1.7663 fringe (-3.72e+20 m^-2), dL = +124.605 um
  (1, 3): dN = -0.6324 fringe (-1.33e+20 m^-2), dL = +17.287 um
  (0, 1): dN = -0.6511 fringe (-1.37e+20 m^-2), dL = +6.895 um
  (1, 1): dN = +0.6698 fringe (+1.41e+20 m^-2), dL = +3.497 um
  (1, 0): dN = +1.3209 fringe (+2.78e+20 m^-2), dL = -3.399 um
  (0, -2): dN = +1.3022 fringe (+2.74e+20 m^-2), dL = -13.791 um
ITER TIP 10.59/4.6
  (1, 2): dN = +0.1618 fringe (+3.41e+19 m^-2), dL = +8.877 um
  (3, 7): dN = -0.0500 fringe (-1.05e+19 m^-2), dL = +32.300 um
  (10, 23): dN = +0.0116 fringe (+2.45e+18 m^-2), dL = +105.777 um
  (1, 3): dN = -0.3736 fringe (-7.87e+19 m^-2), dL = +14.547 um
  (0, 1): dN = -0.5354 fringe (-1.13e+20 m^-2), dL = +5.670 um
  (1, 1): dN = +0.6972 fringe (+1.47e+20 m^-2), dL = +3.207 um
  (1, 0): dN = +1.2326 fringe (+2.60e+20 m^-2), dL = -2.463 um
  (0, -2): dN = +1.0708 fringe (+2.25e+20 m^-2), dL = -11.340 um
density-only boundary for (1,2), prototype: 0.0094 fringe
ITER (3,7) displacement: 0.050047 fringe
prototype, path prior none: d(1,2) = 0.37, wrong count 8.53e-01, branch displacement > 0.05 3.55e-01, total error > 0.05 3.22e-01
prototype, path prior 10 um: d(1,2) = 1.10, wrong count 5.81e-01, branch displacement > 0.05 5.75e-03, total error > 0.05 8.25e-03
prototype, path prior 5 um: d(1,2) = 2.11, wrong count 2.91e-01, branch displacement > 0.05 0.00e+00, total error > 0.05 7.75e-05
   155 total-error events, 95 % interval 6.6e-05 to 9.1e-05; 137 select twice the near-null vector
prototype, path prior 2 um: d(1,2) = 5.21, wrong count 9.18e-03, branch displacement > 0.05 0.00e+00, total error > 0.05 0.00e+00
prototype, path prior 1 um: d(1,2) = 10.39, wrong count 5.00e-07, branch displacement > 0.05 0.00e+00, total error > 0.05 0.00e+00
ITER, path prior none: d(1,2) = 3.20, wrong count 6.20e-01, branch displacement > 0.05 6.20e-01, total error > 0.05 3.39e-01
ITER, path prior 10 um: d(1,2) = 3.32, wrong count 1.79e-01, branch displacement > 0.05 1.79e-01, total error > 0.05 1.35e-01
ITER, path prior 5 um: d(1,2) = 3.66, wrong count 6.91e-02, branch displacement > 0.05 6.91e-02, total error > 0.05 6.87e-02
ITER, path prior 2 um: d(1,2) = 5.48, wrong count 6.21e-03, branch displacement > 0.05 6.21e-03, total error > 0.05 6.21e-03
zero events: one-sided 95 % bound 1.5e-06 for 2e6 draws, 7.5e-06 for 4e5
prototype, path bound 30 um, path prior none: 51 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.854, total error > 0.05 2.03e-02
prototype, path bound 30 um, path prior 5 um: 51 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.291, total error > 0.05 8.50e-05
prototype, path bound 60 um, path prior none: 105 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.852, total error > 0.05 3.22e-01
prototype, path bound 60 um, path prior 5 um: 105 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.292, total error > 0.05 8.75e-05
prototype, path bound 120 um, path prior none: 207 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.854, total error > 0.05 3.22e-01
prototype, path bound 120 um, path prior 5 um: 207 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.292, total error > 0.05 1.00e-04
prototype, path bound 240 um, path prior none: 421 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.853, total error > 0.05 3.22e-01
prototype, path bound 240 um, path prior 5 um: 421 hypotheses, smallest competitor (-1, -2) (-0.0187 fringe, -10.4 um); wrong count 0.290, total error > 0.05 6.75e-05
ITER, path bound 30 um, path prior none: 65 hypotheses, smallest competitor (-1, -2) (-0.1618 fringe, -8.9 um); wrong count 0.110, total error > 0.05 1.10e-01
ITER, path bound 30 um, path prior 5 um: 65 hypotheses, smallest competitor (-1, -2) (-0.1618 fringe, -8.9 um); wrong count 0.069, total error > 0.05 6.88e-02
ITER, path bound 60 um, path prior none: 127 hypotheses, smallest competitor (-3, -7) (+0.0500 fringe, -32.3 um); wrong count 0.620, total error > 0.05 3.39e-01
ITER, path bound 60 um, path prior 5 um: 127 hypotheses, smallest competitor (-3, -7) (+0.0500 fringe, -32.3 um); wrong count 0.069, total error > 0.05 6.85e-02
ITER, path bound 120 um, path prior none: 253 hypotheses, smallest competitor (-10, -23) (-0.0116 fringe, -105.8 um); wrong count 0.908, total error > 0.05 3.84e-01
ITER, path bound 120 um, path prior 5 um: 253 hypotheses, smallest competitor (-10, -23) (-0.0116 fringe, -105.8 um); wrong count 0.069, total error > 0.05 6.81e-02
ITER, path bound 240 um, path prior none: 509 hypotheses, smallest competitor (-10, -23) (-0.0116 fringe, -105.8 um); wrong count 0.908, total error > 0.05 3.19e-01
ITER, path bound 240 um, path prior 5 um: 509 hypotheses, smallest competitor (-10, -23) (-0.0116 fringe, -105.8 um); wrong count 0.069, total error > 0.05 6.82e-02
prototype, path prior 5 um: bound from 1.3e-05 to 2.0e-05
   1 hard-branch estimate, all decisions: 6.6e-05 (132 of 2000000, 95 % interval 5.5e-05 to 7.8e-05)
   2 marginalized estimate, all decisions: 1.7e-05 (33 of 2000000, 95 % interval 1.1e-05 to 2.3e-05), mean bound 1.7e-05
     on the same decisions: 112 errors of the hard-branch estimate only, 13 of the marginalized estimate only
   availability 1.000 (2000000 of 2000000 decisions):
     3 marginalized estimate, bound <= 0.0001: 1.7e-05 (33 of 2000000, 95 % interval 1.1e-05 to 2.3e-05), mean bound 1.7e-05
     4 marginalized estimate, count confidence: 1.7e-05 (33 of 2000000, 95 % interval 1.1e-05 to 2.3e-05); accepted sets differ in 0 decisions
     hard-branch estimate, count confidence: 6.6e-05 (132 of 2000000, 95 % interval 5.5e-05 to 7.8e-05)
   availability 0.819 (1638948 of 2000000 decisions):
     3 marginalized estimate, bound <= 2e-05: 1.5e-05 (24 of 1638948, 95 % interval 9.4e-06 to 2.2e-05), mean bound 1.6e-05
     4 marginalized estimate, count confidence: 1.5e-05 (24 of 1638948, 95 % interval 9.4e-06 to 2.2e-05); accepted sets differ in 0 decisions
     hard-branch estimate, count confidence: 3.8e-05 (63 of 1638948, 95 % interval 3.0e-05 to 4.9e-05)
   count confidence of at least 0.99 in 0 of 2000000 decisions
   reference check on 10000 decisions: bound minus full-mixture risk from -3e-16 to 3e-16; candidate estimate above the reference value by at most 4.2 % (median 2.1 %); withheld by the bound but accepted at the reference value at 0.0001: 0.0 %, withheld by the bound but accepted at the reference value at 2e-05: 4.0 %
ITER, path prior 2 um: bound from 6.1e-07 to 5.0e-01
   1 hard-branch estimate, all decisions: 6.2e-03 (12334 of 2000000, 95 % interval 6.1e-03 to 6.3e-03)
   2 marginalized estimate, all decisions: 6.2e-03 (12334 of 2000000, 95 % interval 6.1e-03 to 6.3e-03), mean bound 6.2e-03
     on the same decisions: 0 errors of the hard-branch estimate only, 0 of the marginalized estimate only
   availability 0.861 (1721259 of 2000000 decisions):
     3 marginalized estimate, bound <= 0.001: 6.3e-05 (108 of 1721259, 95 % interval 5.1e-05 to 7.6e-05), mean bound 7.4e-05
     4 marginalized estimate, count confidence: 6.3e-05 (108 of 1721259, 95 % interval 5.1e-05 to 7.6e-05); accepted sets differ in 0 decisions
     hard-branch estimate, count confidence: 6.3e-05 (108 of 1721259, 95 % interval 5.1e-05 to 7.6e-05)
   availability 0.710 (1419624 of 2000000 decisions):
     3 marginalized estimate, bound <= 0.0001: 1.2e-05 (17 of 1419624, 95 % interval 7.0e-06 to 1.9e-05), mean bound 1.4e-05
     4 marginalized estimate, count confidence: 1.2e-05 (17 of 1419624, 95 % interval 7.0e-06 to 1.9e-05); accepted sets differ in 0 decisions
     hard-branch estimate, count confidence: 1.2e-05 (17 of 1419624, 95 % interval 7.0e-06 to 1.9e-05)
   count confidence of at least 0.99 in 1885490 of 2000000 decisions: hard-branch estimate 3.6e-04 (678 of 1885490, 95 % interval 3.3e-04 to 3.9e-04), marginalized estimate 3.6e-04 (678 of 1885490, 95 % interval 3.3e-04 to 3.9e-04)
   reference check on 10000 decisions: bound minus full-mixture risk from -2e-16 to 2e-16; candidate estimate above the reference value by at most 0.0 % (median 0.0 %); withheld by the bound but accepted at the reference value at 0.001: 0.0 %, withheld by the bound but accepted at the reference value at 0.0001: 0.0 %
(1,2) step: change per nm of lambda2 -1.32 %
lambda2 = 5.22 um: 0.1 um chromatic path difference gives 0.0125 fringe
lambda2 = 4.6 um: 0.1 um chromatic path difference gives 0.0116 fringe
DIII-D prototype 10.59/5.22: 1.9 deg phase noise -> 1.6e+18 m^-2 (0.0078 fringe), 0.041 um, correlation -0.79
ITER TIP 10.59/4.6: 1.9 deg phase noise -> 1.5e+18 m^-2 (0.0071 fringe), 0.033 um, correlation -0.73
C_F = 2.631e-13 rad/T; 1.5e19 T m^-2 at 10.59 um: psi = 0.0254 deg, 2 psi = 0.0507 deg
  mean parallel field 1.5 T: 1.0e+19 m^-2 = 0.047 fringe
  mean parallel field 2.0 T: 7.5e+18 m^-2 = 0.036 fringe
T_e = 25 keV: 3tau/2 = 0.073, 2tau = 0.098, mismatch tau/2 = 0.024; at 1e21 m^-2: 2.4e+19 m^-2 = 0.12 fringe
