{ "results": [ { "name": "historical", "correct": 2895, "n": 3080, "accuracy_percent": 93.9935064935065, "nll": 0.37805166840553284, "seconds": 10.313821391901001 }, { "name": "control_13", "correct": 2895, "n": 3080, "accuracy_percent": 93.9935064935065, "nll": 0.41431134939193726, "seconds": 8.589660748140886 }, { "name": "control_42", "correct": 2894, "n": 3080, "accuracy_percent": 93.96103896103895, "nll": 0.41149818897247314, "seconds": 8.787950488971546 }, { "name": "control_87", "correct": 2892, "n": 3080, "accuracy_percent": 93.8961038961039, "nll": 0.41784676909446716, "seconds": 8.675577850081027 }, { "name": "sam_13", "correct": 2899, "n": 3080, "accuracy_percent": 94.12337662337661, "nll": 0.3676500916481018, "seconds": 8.66914170095697 }, { "name": "sam_42", "correct": 2896, "n": 3080, "accuracy_percent": 94.02597402597402, "nll": 0.37339073419570923, "seconds": 8.70176282688044 }, { "name": "sam_87", "correct": 2894, "n": 3080, "accuracy_percent": 93.96103896103895, "nll": 0.3706142008304596, "seconds": 8.740875865099952 } ], "comparisons": { "control_13": { "delta_pp": 0.0, "shared_row_95ci_pp": [ -0.22727272727272727, 0.22727272727272727 ] }, "control_42": { "delta_pp": -0.032467532467532464, "shared_row_95ci_pp": [ -0.22727272727272727, 0.16233766233766234 ] }, "control_87": { "delta_pp": -0.09740259740259741, "shared_row_95ci_pp": [ -0.35714285714285715, 0.16233766233766234 ] }, "sam_13": { "delta_pp": 0.12987012987012986, "shared_row_95ci_pp": [ -0.16233766233766234, 0.45454545454545453 ] }, "sam_42": { "delta_pp": 0.032467532467532464, "shared_row_95ci_pp": [ -0.2597402597402597, 0.3246753246753247 ] }, "sam_87": { "delta_pp": -0.032467532467532464, "shared_row_95ci_pp": [ -0.3246753246753247, 0.2597402597402597 ] }, "control_mean_vs_historical": { "delta_pp": -0.04329004329004328, "shared_row_95ci_pp": [ -0.24891774891774887, 0.15178571428571036 ] }, "sam_mean_vs_historical": { "delta_pp": 0.04329004329004331, "shared_row_95ci_pp": [ -0.22727272727272727, 0.3246753246753247 ] }, "sam_mean_vs_control_mean": { "delta_pp": 0.0865800865800866, "shared_row_95ci_pp": [ -0.12987012987012986, 0.31385281385281383 ] } }, "scope": "retrospective official test, previously observed; no subsequent tuning" }