{
 "threshold": 0.65,
 "fractions": [
  0.7,
  0.5,
  0.35
 ],
 "primary_f": 0.35,
 "model": {
  "path": "ft44s_best_fp16.onnx",
  "manifest": "/home/agent/work/sieve-test/model/model_manifest.json"
 },
 "sets": {
  "glossy": {
   "rows": 2086,
   "per_condition": {
    "base": {
     "scored": 298,
     "skipped": 0,
     "flagged": 10,
     "rate_pct": 3.3557046979865772,
     "below_440_short_side": 1
    },
    "crop35": {
     "scored": 298,
     "skipped": 0,
     "flagged": 14,
     "rate_pct": 4.697986577181208,
     "below_440_short_side": 71
    },
    "crop50": {
     "scored": 298,
     "skipped": 0,
     "flagged": 15,
     "rate_pct": 5.033557046979865,
     "below_440_short_side": 42
    },
    "crop70": {
     "scored": 298,
     "skipped": 0,
     "flagged": 11,
     "rate_pct": 3.6912751677852347,
     "below_440_short_side": 19
    },
    "shrink35": {
     "scored": 298,
     "skipped": 0,
     "flagged": 11,
     "rate_pct": 3.6912751677852347,
     "below_440_short_side": 71
    },
    "shrink50": {
     "scored": 298,
     "skipped": 0,
     "flagged": 9,
     "rate_pct": 3.0201342281879193,
     "below_440_short_side": 42
    },
    "shrink70": {
     "scored": 297,
     "skipped": 1,
     "flagged": 10,
     "rate_pct": 3.367003367003367,
     "below_440_short_side": 19
    }
   },
   "ladder": {
    "0.70": {
     "n_pairs": 297,
     "a_flagged": 11,
     "b_flagged": 10,
     "a_only": 8,
     "b_only": 7,
     "both": 3,
     "neither": 279,
     "discordant": 15,
     "p": 1.0,
     "a_rate": 3.7037037037037037,
     "b_rate": 3.367003367003367
    },
    "0.50": {
     "n_pairs": 298,
     "a_flagged": 15,
     "b_flagged": 9,
     "a_only": 13,
     "b_only": 7,
     "both": 2,
     "neither": 276,
     "discordant": 20,
     "p": 0.26317596435546875,
     "a_rate": 5.033557046979865,
     "b_rate": 3.0201342281879193
    },
    "0.35": {
     "n_pairs": 298,
     "a_flagged": 14,
     "b_flagged": 11,
     "a_only": 12,
     "b_only": 9,
     "both": 2,
     "neither": 275,
     "discordant": 21,
     "p": 0.6636238098144531,
     "a_rate": 4.697986577181208,
     "b_rate": 3.6912751677852347
    }
   },
   "mean_shift": {
    "0.70": {
     "n": 297,
     "mean_diff": 0.0020188647451470777,
     "sd": 0.17002186307577233,
     "median_diff": 3.78430663624384e-05
    },
    "0.50": {
     "n": 298,
     "mean_diff": 0.022265810420845383,
     "sd": 0.2156028763612396,
     "median_diff": 3.308516743308321e-05
    },
    "0.35": {
     "n": 298,
     "mean_diff": 0.00815606430707203,
     "sd": 0.24041619652938695,
     "median_diff": 0.00019409004740555003
    }
   },
   "b2_base_reproduces_clean": {
    "checked": 298,
    "worst_abs_diff": 0
   },
   "vs_base": {
    "0.70": {
     "n_pairs": 298,
     "a_flagged": 11,
     "b_flagged": 10,
     "a_only": 8,
     "b_only": 7,
     "both": 3,
     "neither": 280,
     "discordant": 15,
     "p": 1.0,
     "a_rate": 3.6912751677852347,
     "b_rate": 3.3557046979865772
    },
    "0.50": {
     "n_pairs": 298,
     "a_flagged": 15,
     "b_flagged": 10,
     "a_only": 13,
     "b_only": 8,
     "both": 2,
     "neither": 275,
     "discordant": 21,
     "p": 0.38331031799316406,
     "a_rate": 5.033557046979865,
     "b_rate": 3.3557046979865772
    },
    "0.35": {
     "n_pairs": 298,
     "a_flagged": 14,
     "b_flagged": 10,
     "a_only": 14,
     "b_only": 10,
     "both": 0,
     "neither": 274,
     "discordant": 24,
     "p": 0.5412561893463135,
     "a_rate": 4.697986577181208,
     "b_rate": 3.3557046979865772
    }
   }
  },
  "control": {
   "rows": 2100,
   "per_condition": {
    "base": {
     "scored": 300,
     "skipped": 0,
     "flagged": 9,
     "rate_pct": 3.0,
     "below_440_short_side": 1
    },
    "crop35": {
     "scored": 299,
     "skipped": 1,
     "flagged": 13,
     "rate_pct": 4.3478260869565215,
     "below_440_short_side": 32
    },
    "crop50": {
     "scored": 299,
     "skipped": 1,
     "flagged": 15,
     "rate_pct": 5.016722408026756,
     "below_440_short_side": 22
    },
    "crop70": {
     "scored": 300,
     "skipped": 0,
     "flagged": 12,
     "rate_pct": 4.0,
     "below_440_short_side": 11
    },
    "shrink35": {
     "scored": 300,
     "skipped": 0,
     "flagged": 7,
     "rate_pct": 2.3333333333333335,
     "below_440_short_side": 32
    },
    "shrink50": {
     "scored": 300,
     "skipped": 0,
     "flagged": 7,
     "rate_pct": 2.3333333333333335,
     "below_440_short_side": 22
    },
    "shrink70": {
     "scored": 300,
     "skipped": 0,
     "flagged": 7,
     "rate_pct": 2.3333333333333335,
     "below_440_short_side": 11
    }
   },
   "ladder": {
    "0.70": {
     "n_pairs": 300,
     "a_flagged": 12,
     "b_flagged": 7,
     "a_only": 7,
     "b_only": 2,
     "both": 5,
     "neither": 286,
     "discordant": 9,
     "p": 0.1796875,
     "a_rate": 4.0,
     "b_rate": 2.3333333333333335
    },
    "0.50": {
     "n_pairs": 299,
     "a_flagged": 15,
     "b_flagged": 7,
     "a_only": 12,
     "b_only": 4,
     "both": 3,
     "neither": 280,
     "discordant": 16,
     "p": 0.076812744140625,
     "a_rate": 5.016722408026756,
     "b_rate": 2.3411371237458196
    },
    "0.35": {
     "n_pairs": 299,
     "a_flagged": 13,
     "b_flagged": 7,
     "a_only": 11,
     "b_only": 5,
     "both": 2,
     "neither": 281,
     "discordant": 16,
     "p": 0.210113525390625,
     "a_rate": 4.3478260869565215,
     "b_rate": 2.3411371237458196
    }
   },
   "mean_shift": {
    "0.70": {
     "n": 300,
     "mean_diff": 0.009347419076874553,
     "sd": 0.1417322835620846,
     "median_diff": -2.1496573508886674e-05
    },
    "0.50": {
     "n": 299,
     "mean_diff": 0.009645177249425249,
     "sd": 0.19273596426442313,
     "median_diff": -0.00026979316055465236
    },
    "0.35": {
     "n": 299,
     "mean_diff": 0.022505189219226618,
     "sd": 0.20232581022590787,
     "median_diff": 1.500016503708213e-06
    }
   },
   "b2_base_reproduces_clean": {
    "checked": 300,
    "worst_abs_diff": 0
   },
   "vs_base": {
    "0.70": {
     "n_pairs": 300,
     "a_flagged": 12,
     "b_flagged": 9,
     "a_only": 7,
     "b_only": 4,
     "both": 5,
     "neither": 284,
     "discordant": 11,
     "p": 0.548828125,
     "a_rate": 4.0,
     "b_rate": 3.0
    },
    "0.50": {
     "n_pairs": 299,
     "a_flagged": 15,
     "b_flagged": 8,
     "a_only": 13,
     "b_only": 6,
     "both": 2,
     "neither": 278,
     "discordant": 19,
     "p": 0.1670684814453125,
     "a_rate": 5.016722408026756,
     "b_rate": 2.6755852842809364
    },
    "0.35": {
     "n_pairs": 299,
     "a_flagged": 13,
     "b_flagged": 8,
     "a_only": 11,
     "b_only": 6,
     "both": 2,
     "neither": 280,
     "discordant": 17,
     "p": 0.332305908203125,
     "a_rate": 4.3478260869565215,
     "b_rate": 2.6755852842809364
    }
   }
  }
 },
 "primary": {
  "comparison": "crop35 vs shrink35 on glossy",
  "n_pairs": 298,
  "a_flagged": 14,
  "b_flagged": 11,
  "a_only": 12,
  "b_only": 9,
  "both": 2,
  "neither": 275,
  "discordant": 21,
  "p": 0.6636238098144531,
  "a_rate": 4.697986577181208,
  "b_rate": 3.6912751677852347,
  "verdict": "H1 NOT SUPPORTED: crop 14/298 vs shrink 11/298, discordant 12/9, p=0.6636"
 }
}