{
 "branch": "pretrain",
 "within_group_clause": {
  "status": "UNTESTABLE",
  "why": "the clause compares members' STABILITY dates, and A38's spacing check leaves no construct of P1 with one on this branch: 0 dated in the product space and 0 in the verdict layer once two spacing bands must agree",
  "n_dated_product": 0,
  "n_dated_layer": 0,
  "spacing_dependence_spearman": -0.8131468225548475
 },
 "exploratory_level_and_trend": {
  "for": {
   "members": [
    "AsyncFor",
    "DictComp",
    "For",
    "GeneratorExp",
    "ListComp",
    "SetComp"
   ],
   "levels": {
    "AsyncFor": 0.4846707177493174,
    "DictComp": 0.5528684562618681,
    "For": 0.5877919141737398,
    "GeneratorExp": 0.5607898813269776,
    "ListComp": 0.5651314580249929,
    "SetComp": 0.5580844744638028
   },
   "trends": {
    "AsyncFor": [
     -0.5080213903743316,
     0.0027
    ],
    "DictComp": [
     -0.607620320855615,
     0.0006
    ],
    "For": [
     -0.4602272727272727,
     0.0064
    ],
    "GeneratorExp": [
     -0.6092914438502673,
     0.0004
    ],
    "ListComp": [
     -0.5471256684491979,
     0.0014
    ],
    "SetComp": [
     -0.5932486631016043,
     0.0005
    ]
   },
   "norm_dates": {
    "AsyncFor": 1908408320000,
    "DictComp": 977272832000,
    "GeneratorExp": 499122176000,
    "ListComp": 33554432000,
    "SetComp": 624951296000
   },
   "role_frequency_per_file": {
    "AsyncFor": 0.0014404217870701427,
    "DictComp": 0.054230893692925754,
    "For": 2.7936526728703464,
    "GeneratorExp": 0.22663162298044973,
    "ListComp": 0.8200459356427439,
    "SetComp": 0.011752263238068177
   },
   "level_spread": 0.10312119642442236,
   "levels_within_one_noise_unit": false,
   "norm_dates_coincide": false,
   "norm_date_spread_decades": 1.7549214096651689,
   "spearman_date_vs_frequency": -0.8225427863778353
  },
  "with": {
   "members": [
    "AsyncWith",
    "With"
   ],
   "levels": {
    "AsyncWith": 0.47149241117231333,
    "With": 0.5116013034704975
   },
   "trends": {
    "AsyncWith": [
     -0.5858957219251337,
     0.0005
    ],
    "With": [
     -0.6026069518716577,
     0.0003
    ]
   },
   "norm_dates": {},
   "role_frequency_per_file": {
    "AsyncWith": 0.003942414699405678,
    "With": 1.0171232606413625
   },
   "level_spread": 0.04010889229818415,
   "levels_within_one_noise_unit": false
  },
  "if": {
   "members": [
    "If",
    "IfExp"
   ],
   "levels": {
    "If": 0.592573118868545,
    "IfExp": 0.5790821037151995
   },
   "trends": {
    "If": [
     -0.4535427807486631,
     0.0087
    ],
    "IfExp": [
     -0.5745320855614974,
     0.0008
    ]
   },
   "norm_dates": {
    "IfExp": 1526726656000
   },
   "role_frequency_per_file": {
    "If": 10.767073931128106,
    "IfExp": 0.37722870740889825
   },
   "level_spread": 0.013491015153345498,
   "levels_within_one_noise_unit": true
  }
 },
 "exploratory_norm_dates": {},
 "exploratory_two_predictor": {
  "with_pilot|lognormal": {
   "n": 162,
   "n_events": 73,
   "models": {
    "both": {
     "predictors": [
      "log_frequency",
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -547.256517906363,
     "coefficients": {
      "intercept": {
       "beta": 27.452235058232027,
       "lo": 26.888744159200733,
       "hi": 27.906751640158838,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.3261459125993451,
       "lo": -0.4828114056205755,
       "hi": -0.21853223599086474,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": 0.030549133816755142,
       "lo": -0.36644300354099607,
       "hi": 0.6398129257140309,
       "excludes_zero": false
      }
     },
     "n_bootstrap": 200
    },
    "frequency_alone": {
     "predictors": [
      "log_frequency"
     ],
     "converged": true,
     "log_likelihood": -547.2689294905076,
     "coefficients": {
      "intercept": {
       "beta": 27.4698687356487,
       "lo": 27.053663837718,
       "hi": 27.899301423222123,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.32354409186761746,
       "lo": -0.4468756888215359,
       "hi": -0.20720410578129608,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    },
    "exposure_alone": {
     "predictors": [
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -566.7040175938164,
     "coefficients": {
      "intercept": {
       "beta": 28.744101915668807,
       "lo": 28.43571082305455,
       "hi": 29.082685453270177,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": -0.4207998476464216,
       "lo": -0.819690215804176,
       "hi": -0.01814055908952038,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    }
   },
   "exposure_total_coefficient": {
    "value": 0.030549133816755142,
    "note": "in the (log frequency, log multiplier) parameterisation the multiplier's coefficient IS the coefficient on log total exposure holding frequency fixed, since log exposure = log frequency + log multiplier"
   }
  },
  "with_pilot|weibull": {
   "n": 162,
   "n_events": 73,
   "models": {
    "both": {
     "predictors": [
      "log_frequency",
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -544.7742205347744,
     "coefficients": {
      "intercept": {
       "beta": 27.964159727961206,
       "lo": 27.548494530140953,
       "hi": 28.309324423281183,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.243489861227487,
       "lo": -0.3913851785189969,
       "hi": -0.15439659406482714,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": 0.11926604484257766,
       "lo": -0.17169139747501083,
       "hi": 0.5580170539613779,
       "excludes_zero": false
      }
     },
     "n_bootstrap": 200
    },
    "frequency_alone": {
     "predictors": [
      "log_frequency"
     ],
     "converged": true,
     "log_likelihood": -545.0650239551546,
     "coefficients": {
      "intercept": {
       "beta": 28.05694070199219,
       "lo": 27.800171961875062,
       "hi": 28.29547132894031,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.22552302065350452,
       "lo": -0.316517099045804,
       "hi": -0.1473092469049648,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    },
    "exposure_alone": {
     "predictors": [
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -561.0559614677179,
     "coefficients": {
      "intercept": {
       "beta": 28.93062939793549,
       "lo": 28.702876137135586,
       "hi": 29.270869209763056,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": -0.3780833837070209,
       "lo": -0.6104188056988531,
       "hi": -0.13188371177148822,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    }
   },
   "exposure_total_coefficient": {
    "value": 0.11926604484257766,
    "note": "in the (log frequency, log multiplier) parameterisation the multiplier's coefficient IS the coefficient on log total exposure holding frequency fixed, since log exposure = log frequency + log multiplier"
   }
  },
  "without_pilot|lognormal": {
   "n": 149,
   "n_events": 65,
   "models": {
    "both": {
     "predictors": [
      "log_frequency",
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -483.16782717370194,
     "coefficients": {
      "intercept": {
       "beta": 27.16920168461071,
       "lo": 26.575790567551035,
       "hi": 27.602113936379176,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.3874442116103855,
       "lo": -0.5727239900399403,
       "hi": -0.2591748080795826,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": 0.100734752048378,
       "lo": -0.3051406983080099,
       "hi": 0.868479948845614,
       "excludes_zero": false
      }
     },
     "n_bootstrap": 200
    },
    "frequency_alone": {
     "predictors": [
      "log_frequency"
     ],
     "converged": true,
     "log_likelihood": -483.2448968290138,
     "coefficients": {
      "intercept": {
       "beta": 27.193038467240115,
       "lo": 26.755310696829177,
       "hi": 27.608774987677844,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.38547785880080515,
       "lo": -0.5092048502531483,
       "hi": -0.24924767137831835,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    },
    "exposure_alone": {
     "predictors": [
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -508.6728911020474,
     "coefficients": {
      "intercept": {
       "beta": 28.71285334233304,
       "lo": 28.441944483373714,
       "hi": 29.05100647364153,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": -0.24144778294660219,
       "lo": -0.5311939066506085,
       "hi": 0.12028595775034973,
       "excludes_zero": false
      }
     },
     "n_bootstrap": 200
    }
   },
   "exposure_total_coefficient": {
    "value": 0.100734752048378,
    "note": "in the (log frequency, log multiplier) parameterisation the multiplier's coefficient IS the coefficient on log total exposure holding frequency fixed, since log exposure = log frequency + log multiplier"
   }
  },
  "without_pilot|weibull": {
   "n": 149,
   "n_events": 65,
   "models": {
    "both": {
     "predictors": [
      "log_frequency",
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -481.1503656103173,
     "coefficients": {
      "intercept": {
       "beta": 27.769151920942797,
       "lo": 27.37620154862592,
       "hi": 28.107416203105704,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.29045977610707574,
       "lo": -0.43460058385626643,
       "hi": -0.19034008625411813,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": 0.11341103175916134,
       "lo": -0.27835982028427764,
       "hi": 0.5019005757138669,
       "excludes_zero": false
      }
     },
     "n_bootstrap": 200
    },
    "frequency_alone": {
     "predictors": [
      "log_frequency"
     ],
     "converged": true,
     "log_likelihood": -481.32933083719195,
     "coefficients": {
      "intercept": {
       "beta": 27.816491642958844,
       "lo": 27.486945549471585,
       "hi": 28.06628372139525,
       "excludes_zero": true
      },
      "log_frequency": {
       "beta": -0.28195677314282863,
       "lo": -0.39246646042161204,
       "hi": -0.17932301114977048,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    },
    "exposure_alone": {
     "predictors": [
      "log_exposure_multiplier"
     ],
     "converged": true,
     "log_likelihood": -503.76743882475506,
     "coefficients": {
      "intercept": {
       "beta": 28.900026808597318,
       "lo": 28.676437873786966,
       "hi": 29.204911535795745,
       "excludes_zero": true
      },
      "log_exposure_multiplier": {
       "beta": -0.3414021405939167,
       "lo": -0.6136916516554873,
       "hi": -0.04623076651821903,
       "excludes_zero": true
      }
     },
     "n_bootstrap": 200
    }
   },
   "exposure_total_coefficient": {
    "value": 0.11341103175916134,
    "note": "in the (log frequency, log multiplier) parameterisation the multiplier's coefficient IS the coefficient on log total exposure holding frequency fixed, since log exposure = log frequency + log multiplier"
   }
  }
 },
 "marked": "both forms below are EXPLORATORY forms of P3 and are never its verdict (the Settle chat, 15 Oct)"
}