========================================================================================================================
SUITE4 EXPERIMENT 2 — COUNCIL OF SPECIALISTS + COMPETENCE MODEL
V2-COMPATIBLE COMPLETE OPERATIONAL EVALUATION ENABLED
========================================================================================================================

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=1
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4397/0.2941/0.3210/0.0539
  block=54 n=2,558 AP q/global/recent/rel=0.7797/0.6856/0.6075/0.4187
  block=56 n=3,349 AP q/global/recent/rel=0.4478/0.3421/0.3751/0.4108
  block=57 n=3,201 AP q/global/recent/rel=0.3318/0.3237/0.3270/0.1261
  block=59 n=3,170 AP q/global/recent/rel=0.1523/0.0260/0.0158/0.0038
  block=60 n=2,945 AP q/global/recent/rel=0.1989/0.0178/0.1373/0.1525
[POLICY] alpha=0.00 temperature=0.50
committee                    AP/AUC=0.5024/0.9301 F2@20%=0.1101 P/R=0.0246/0.8235
competence_council           AP/AUC=0.5024/0.9301 F2@20%=0.1101 P/R=0.0246/0.8235 ΔAP=+0.0000
[COMPETENCE] target top1 accuracy=0.720 | NON_DEPLOYABLE best-expert oracle AP=0.9364
  history0     n=1,100 AP 0.1733->0.1733 weights=[0.271 0.264 0.247 0.218]
  history1_2   n=880 AP 0.0481->0.0481 weights=[0.263 0.258 0.248 0.231]
  history3plus n=860 AP 0.6754->0.6754 weights=[0.267 0.262 0.254 0.217]

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=2
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0, 62.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4250/0.3557/0.2875/0.0163
  block=54 n=2,558 AP q/global/recent/rel=0.7744/0.7740/0.7119/0.4460
  block=56 n=3,349 AP q/global/recent/rel=0.4490/0.3057/0.3227/0.3774
  block=57 n=3,201 AP q/global/recent/rel=0.3392/0.3451/0.3336/0.1232
  block=59 n=3,170 AP q/global/recent/rel=0.1529/0.1468/0.0401/0.0239
  block=60 n=2,945 AP q/global/recent/rel=0.1878/0.2621/0.3053/0.0793
  block=62 n=2,840 AP q/global/recent/rel=0.5291/0.1907/0.4008/0.1647
[POLICY] alpha=0.25 temperature=2.00
committee                    AP/AUC=0.5422/0.9638 F2@20%=0.1030 P/R=0.0224/1.0000
competence_council           AP/AUC=0.5463/0.9644 F2@20%=0.1030 P/R=0.0224/1.0000 ΔAP=+0.0041
[COMPETENCE] target top1 accuracy=0.930 | NON_DEPLOYABLE best-expert oracle AP=0.9805
  history0     n=4 AP nan->nan weights=[0.35  0.138 0.224 0.289]
  history1_2   n=928 AP 0.0251->0.0316 weights=[0.266 0.264 0.257 0.213]
  history3plus n=1,516 AP 0.6545->0.6558 weights=[0.267 0.265 0.259 0.209]

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=3
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0, 62.0, 63.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4102/0.3364/0.2913/0.1069
  block=54 n=2,558 AP q/global/recent/rel=0.7735/0.7486/0.7693/0.4964
  block=56 n=3,349 AP q/global/recent/rel=0.4686/0.3859/0.3812/0.3783
  block=57 n=3,201 AP q/global/recent/rel=0.3330/0.3838/0.3293/0.1344
  block=59 n=3,170 AP q/global/recent/rel=0.1534/0.0341/0.0472/0.1551
  block=60 n=2,945 AP q/global/recent/rel=0.2171/0.0888/0.1514/0.0353
  block=62 n=2,840 AP q/global/recent/rel=0.4631/0.0726/0.0969/0.0972
  block=63 n=2,448 AP q/global/recent/rel=0.5395/0.2162/0.3042/0.1275
[POLICY] alpha=0.00 temperature=0.50
committee                    AP/AUC=0.3773/0.9403 F2@20%=0.1365 P/R=0.0309/0.9412
competence_council           AP/AUC=0.3773/0.9403 F2@20%=0.1365 P/R=0.0309/0.9412 ΔAP=+0.0000
[COMPETENCE] target top1 accuracy=0.708 | NON_DEPLOYABLE best-expert oracle AP=1.0000
  history0     n=1,009 AP 0.2204->0.2204 weights=[0.322 0.301 0.279 0.099]
  history1_2   n=956 AP 0.1449->0.1449 weights=[0.32  0.31  0.266 0.104]
  history3plus n=623 AP 0.8385->0.8385 weights=[0.317 0.309 0.278 0.096]

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=4
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0, 62.0, 63.0, 65.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4264/0.3189/0.3264/0.3403
  block=54 n=2,558 AP q/global/recent/rel=0.7571/0.7433/0.7153/0.4660
  block=56 n=3,349 AP q/global/recent/rel=0.4208/0.3414/0.3671/0.4998
  block=57 n=3,201 AP q/global/recent/rel=0.4181/0.3266/0.3945/0.1226
  block=59 n=3,170 AP q/global/recent/rel=0.1527/0.1555/0.1566/0.1581
  block=60 n=2,945 AP q/global/recent/rel=0.2057/0.0567/0.3279/0.1319
  block=62 n=2,840 AP q/global/recent/rel=0.4935/0.0098/0.0798/0.1364
  block=63 n=2,448 AP q/global/recent/rel=0.5792/0.2333/0.4068/0.0988
  block=65 n=2,588 AP q/global/recent/rel=0.3398/0.3287/0.0610/0.1231
[POLICY] alpha=0.25 temperature=0.50
committee                    AP/AUC=0.1667/0.8613 F2@20%=0.2205 P/R=0.0580/0.7368
competence_council           AP/AUC=0.1653/0.8052 F2@20%=0.2362 P/R=0.0621/0.7895 ΔAP=-0.0015
[COMPETENCE] target top1 accuracy=0.837 | NON_DEPLOYABLE best-expert oracle AP=0.8710
  history0     n=4 AP nan->nan weights=[0.299 0.289 0.298 0.113]
  history1_2   n=877 AP 0.1317->0.1317 weights=[0.309 0.298 0.298 0.096]
  history3plus n=1,532 AP 0.2206->0.2181 weights=[0.304 0.295 0.3   0.102]

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=5
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0, 62.0, 63.0, 65.0, 66.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4320/0.2895/0.3584/0.1284
  block=54 n=2,558 AP q/global/recent/rel=0.7791/0.7236/0.6010/0.5454
  block=56 n=3,349 AP q/global/recent/rel=0.4281/0.2283/0.3020/0.4726
  block=57 n=3,201 AP q/global/recent/rel=0.4024/0.3562/0.3335/0.1073
  block=59 n=3,170 AP q/global/recent/rel=0.1540/0.1479/0.1465/0.0330
  block=60 n=2,945 AP q/global/recent/rel=0.2067/0.0536/0.0917/0.1991
  block=62 n=2,840 AP q/global/recent/rel=0.5219/0.0526/0.2972/0.6044
  block=63 n=2,448 AP q/global/recent/rel=0.4984/0.2729/0.3367/0.2127
  block=65 n=2,588 AP q/global/recent/rel=0.3015/0.2693/0.0377/0.0601
  block=66 n=2,413 AP q/global/recent/rel=0.1663/0.1375/0.1263/0.1692
[POLICY] alpha=0.25 temperature=2.00
committee                    AP/AUC=0.2264/0.8362 F2@20%=0.3333 P/R=0.1021/0.7681
competence_council           AP/AUC=0.2265/0.8476 F2@20%=0.3333 P/R=0.1021/0.7681 ΔAP=+0.0001
[COMPETENCE] target top1 accuracy=0.659 | NON_DEPLOYABLE best-expert oracle AP=0.9219
  history0     n=749 AP 0.3474->0.3230 weights=[0.262 0.253 0.258 0.227]
  history1_2   n=1,207 AP 0.3164->0.3156 weights=[0.258 0.249 0.252 0.241]
  history3plus n=637 AP 0.2213->0.2416 weights=[0.255 0.252 0.252 0.24 ]

========================================================================================================================
DEVELOPMENT-SELECTED OPERATIONAL POLICIES [beta=2]
Thresholds and warning ratios below are frozen before holdout evaluation.
========================================================================================================================
  competence_council  : AP=0.2057 | AUC=0.8714 | F2-opt tau=0.124482 | P=0.1586 R=0.3882 F1=0.2252 F2=0.3010 | flag=2.89% | ratio=2.447
  committee           : AP=0.2158 | AUC=0.8821 | F2-opt tau=0.107967 | P=0.1731 R=0.4145 F1=0.2442 F2=0.3241 | flag=2.83% | ratio=2.395

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=6
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0, 62.0, 63.0, 65.0, 66.0, 68.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4402/0.2989/0.3326/0.3876
  block=54 n=2,558 AP q/global/recent/rel=0.7756/0.6953/0.7625/0.5205
  block=56 n=3,349 AP q/global/recent/rel=0.4312/0.4387/0.3364/0.4268
  block=57 n=3,201 AP q/global/recent/rel=0.3330/0.3431/0.3232/0.1393
  block=59 n=3,170 AP q/global/recent/rel=0.1534/0.0286/0.1550/0.0537
  block=60 n=2,945 AP q/global/recent/rel=0.2135/0.0314/0.0195/0.0295
  block=62 n=2,840 AP q/global/recent/rel=0.5346/0.3830/0.3952/0.1591
  block=63 n=2,448 AP q/global/recent/rel=0.5259/0.3318/0.4208/0.2384
  block=65 n=2,588 AP q/global/recent/rel=0.2922/0.2215/0.0505/0.0985
  block=66 n=2,413 AP q/global/recent/rel=0.1642/0.1426/0.1232/0.1914
  block=68 n=2,593 AP q/global/recent/rel=0.2052/0.2336/0.1736/0.1459
[POLICY] alpha=0.00 temperature=0.50
committee                    AP/AUC=0.1763/0.8325 F2@20%=0.3322 P/R=0.1126/0.6477
competence_council           AP/AUC=0.1763/0.8325 F2@20%=0.3322 P/R=0.1126/0.6477 ΔAP=+0.0000
[COMPETENCE] target top1 accuracy=0.856 | NON_DEPLOYABLE best-expert oracle AP=0.8584
  history0     n=7 AP 0.2429->0.2429 weights=[0.462 0.231 0.144 0.163]
  history1_2   n=664 AP 0.2746->0.2746 weights=[0.312 0.278 0.282 0.128]
  history3plus n=1,859 AP 0.1511->0.1511 weights=[0.31  0.278 0.274 0.138]

========================================================================================================================
EXP2 COMPETENCE COUNCIL | fold=7
[FORWARD EXPERTS] blocks=[53.0, 54.0, 56.0, 57.0, 59.0, 60.0, 62.0, 63.0, 65.0, 66.0, 68.0, 69.0]
  block=53 n=2,785 AP q/global/recent/rel=0.4315/0.3440/0.3381/0.3759
  block=54 n=2,558 AP q/global/recent/rel=0.7784/0.6406/0.7579/0.4545
  block=56 n=3,349 AP q/global/recent/rel=0.4465/0.3706/0.3932/0.3550
  block=57 n=3,201 AP q/global/recent/rel=0.3450/0.3408/0.3641/0.1030
  block=59 n=3,170 AP q/global/recent/rel=0.1548/0.0327/0.1660/0.0759
  block=60 n=2,945 AP q/global/recent/rel=0.2043/0.0173/0.0566/0.0063
  block=62 n=2,840 AP q/global/recent/rel=0.5286/0.4742/0.0266/0.1700
  block=63 n=2,448 AP q/global/recent/rel=0.5442/0.2073/0.3202/0.3595
  block=65 n=2,588 AP q/global/recent/rel=0.2936/0.3362/0.0725/0.1155
  block=66 n=2,413 AP q/global/recent/rel=0.1661/0.1534/0.1251/0.2247
  block=68 n=2,593 AP q/global/recent/rel=0.2033/0.1842/0.0774/0.1630
  block=69 n=2,530 AP q/global/recent/rel=0.1865/0.1742/0.1472/0.1161
[POLICY] alpha=0.75 temperature=1.00
committee                    AP/AUC=0.2195/0.8187 F2@20%=0.4170 P/R=0.1699/0.6552
competence_council           AP/AUC=0.2626/0.8295 F2@20%=0.4478 P/R=0.1825/0.7034 ΔAP=+0.0431
[COMPETENCE] target top1 accuracy=0.583 | NON_DEPLOYABLE best-expert oracle AP=0.9224
  history0     n=706 AP 0.3157->0.3071 weights=[0.276 0.257 0.263 0.205]
  history1_2   n=1,430 AP 0.1775->0.2306 weights=[0.274 0.254 0.259 0.213]
  history3plus n=659 AP 0.2159->0.2653 weights=[0.255 0.264 0.279 0.202]

========================================================================================================================
SUITE4 ORIGINAL RANKING / EQUAL-BUDGET SUMMARY
========================================================================================================================
development NON_DEPLOYABLE_best_expert_oracle  pooled AP=0.9118 mean-fold AP=0.9420 F2@20=0.2230
development committee                          pooled AP=0.2158 mean-fold AP=0.3630 F2@20=0.1807
development competence_council                 pooled AP=0.2057 mean-fold AP=0.3636 F2@20=0.1838
holdout     NON_DEPLOYABLE_best_expert_oracle  pooled AP=0.8971 mean-fold AP=0.8904 F2@20=0.5725
holdout     committee                          pooled AP=0.1915 mean-fold AP=0.1979 F2@20=0.3746
holdout     competence_council                 pooled AP=0.2127 mean-fold AP=0.2195 F2@20=0.3900

### DEVELOPMENT [model=competence_council, beta=2]
  fold    test   base% |      AP     AUC |     P@t     R@t    F1@t    Fβ@t   rate@t |     P@r     R@r    F1@r    Fβ@r   rate@r
     1    2840   0.60% |  0.5024  0.9301 |  1.0000  0.2941  0.4545  0.3425    0.18% |  0.1098  0.5294  0.1818  0.3000    2.89%
     2    2448   0.45% |  0.5463  0.9644 |  0.2800  0.6364  0.3889  0.5072    1.02% |  0.1944  0.6364  0.2979  0.4375    1.47%
     3    2588   0.66% |  0.3773  0.9403 |  0.2414  0.4118  0.3043  0.3608    1.12% |  0.2143  0.3529  0.2667  0.3125    1.08%
     4    2413   1.57% |  0.1653  0.8052 |  0.1429  0.2105  0.1702  0.1923    2.32% |  0.1282  0.1316  0.1299  0.1309    1.62%
     5    2593   2.66% |  0.2265  0.8476 |  0.1245  0.4638  0.1963  0.3002    9.91% |  0.1800  0.2609  0.2130  0.2394    3.86%
------------------------------------------------------------------------------------------------------------------------------------
POOLED   12882   1.18% |  0.2057  0.8714 |  0.1586  0.3882  0.2252  0.3010    2.89% |  0.1579  0.2961  0.2059  0.2520    2.21%

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
HOLDOUT: pooled development threshold and warning ratio are frozen. No holdout label is used for threshold tuning.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

### HOLDOUT (test) [model=competence_council, beta=2]
  fold    test   base% |      AP     AUC |     P@t     R@t    F1@t    Fβ@t   rate@t |     P@r     R@r    F1@r    Fβ@r   rate@r
     6    2530   3.48% |  0.1763  0.8325 |  0.2000  0.4545  0.2778  0.3623    7.91% |  0.2242  0.4205  0.2925  0.3578    6.52%

  Confusion matrix — fold 6 — fixed development threshold 0.124482
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,282             160
  Failure <5                            48              40
  ------------------------------------------------------------------
  Normal correctly rejected: 93.45%
  Failures successfully detected: 45.45%

  Confusion matrix — fold 6 — drift-aware rate 6.512%
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,314             128
  Failure <5                            51              37
  ------------------------------------------------------------------
  Normal correctly rejected: 94.76%
  Failures successfully detected: 42.05%
     7    2795   5.19% |  0.2626  0.8295 |  0.1520  0.7862  0.2547  0.4286   26.83% |  0.2647  0.4345  0.3290  0.3851    8.52%

  Confusion matrix — fold 7 — fixed development threshold 0.124482
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,014             636
  Failure <5                            31             114
  ------------------------------------------------------------------
  Normal correctly rejected: 76.00%
  Failures successfully detected: 78.62%

  Confusion matrix — fold 7 — drift-aware rate 8.513%
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,475             175
  Failure <5                            82              63
  ------------------------------------------------------------------
  Normal correctly rejected: 93.40%
  Failures successfully detected: 43.45%
------------------------------------------------------------------------------------------------------------------------------------
POOLED    5325   4.38% |  0.2127  0.8009 |  0.1621  0.6609  0.2604  0.4091   17.84% |  0.2481  0.4292  0.3145  0.3745    7.57%

  Confusion matrix — POOLED — fixed development threshold
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         4,296             796
  Failure <5                            79             154
  ------------------------------------------------------------------
  Normal correctly rejected: 84.37%
  Failures successfully detected: 66.09%

  Confusion matrix — POOLED — drift-aware warning policy
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         4,789             303
  Failure <5                           133             100
  ------------------------------------------------------------------
  Normal correctly rejected: 94.05%
  Failures successfully detected: 42.92%

### DEVELOPMENT [model=committee, beta=2]
  fold    test   base% |      AP     AUC |     P@t     R@t    F1@t    Fβ@t   rate@t |     P@r     R@r    F1@r    Fβ@r   rate@r
     1    2840   0.60% |  0.5024  0.9301 |  0.8750  0.4118  0.5600  0.4605    0.28% |  0.1125  0.5294  0.1856  0.3041    2.82%
     2    2448   0.45% |  0.5422  0.9638 |  0.2593  0.6364  0.3684  0.4930    1.10% |  0.2000  0.6364  0.3043  0.4430    1.43%
     3    2588   0.66% |  0.3773  0.9403 |  0.2128  0.5882  0.3125  0.4348    1.82% |  0.2143  0.3529  0.2667  0.3125    1.08%
     4    2413   1.57% |  0.1667  0.8613 |  0.1404  0.2105  0.1684  0.1914    2.36% |  0.1316  0.1316  0.1316  0.1316    1.57%
     5    2593   2.66% |  0.2264  0.8362 |  0.1378  0.4493  0.2109  0.3094    8.68% |  0.1735  0.2464  0.2036  0.2273    3.78%
------------------------------------------------------------------------------------------------------------------------------------
POOLED   12882   1.18% |  0.2158  0.8821 |  0.1731  0.4145  0.2442  0.3241    2.83% |  0.1577  0.2895  0.2042  0.2480    2.17%

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
HOLDOUT: pooled development threshold and warning ratio are frozen. No holdout label is used for threshold tuning.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

### HOLDOUT (test) [model=committee, beta=2]
  fold    test   base% |      AP     AUC |     P@t     R@t    F1@t    Fβ@t   rate@t |     P@r     R@r    F1@r    Fβ@r   rate@r
     6    2530   3.48% |  0.1763  0.8325 |  0.2087  0.4886  0.2925  0.3853    8.14% |  0.2298  0.4205  0.2972  0.3606    6.36%

  Confusion matrix — fold 6 — fixed development threshold 0.107967
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,279             163
  Failure <5                            45              43
  ------------------------------------------------------------------
  Normal correctly rejected: 93.33%
  Failures successfully detected: 48.86%

  Confusion matrix — fold 6 — drift-aware rate 6.372%
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,318             124
  Failure <5                            51              37
  ------------------------------------------------------------------
  Normal correctly rejected: 94.92%
  Failures successfully detected: 42.05%
     7    2795   5.19% |  0.2195  0.8187 |  0.2051  0.5034  0.2914  0.3900   12.74% |  0.2446  0.3931  0.3016  0.3506    8.34%

  Confusion matrix — fold 7 — fixed development threshold 0.107967
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,367             283
  Failure <5                            72              73
  ------------------------------------------------------------------
  Normal correctly rejected: 89.32%
  Failures successfully detected: 50.34%

  Confusion matrix — fold 7 — drift-aware rate 8.330%
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         2,474             176
  Failure <5                            88              57
  ------------------------------------------------------------------
  Normal correctly rejected: 93.36%
  Failures successfully detected: 39.31%
------------------------------------------------------------------------------------------------------------------------------------
POOLED    5325   4.38% |  0.1915  0.8156 |  0.2064  0.4979  0.2918  0.3882   10.55% |  0.2386  0.4034  0.2998  0.3544    7.40%

  Confusion matrix — POOLED — fixed development threshold
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         4,646             446
  Failure <5                           117             116
  ------------------------------------------------------------------
  Normal correctly rejected: 91.24%
  Failures successfully detected: 49.79%

  Confusion matrix — POOLED — drift-aware warning policy
  ------------------------------------------------------------------
  Actual / Predicted            Normal >=5      Failure <5
  ------------------------------------------------------------------
  Normal >=5                         4,792             300
  Failure <5                           139              94
  ------------------------------------------------------------------
  Normal correctly rejected: 94.11%
  Failures successfully detected: 40.34%
