{
 "slug": "consensus-watch-2026-08-31",
 "series": "weekly",
 "series_label": "WEEKLY · CONSENSUS",
 "issue": 5,
 "title": "Consensus Watch #5",
 "language": "en",
 "window_start": "2026-08-31T00:00:00+00:00",
 "window_end": "2026-09-07T00:00:00+00:00",
 "window_text": "Aug 31-Sep 6, 2026 UTC",
 "cutoff": "2026-09-07T16:00:00+00:00",
 "generated_at": "2026-09-08T09:12:03+00:00",
 "methodology": "v1.1 (2026-08-10)",
 "methodology_hash": "e66c7e8c864a2233",
 "source": "weekly_metrics_2026-08-31.json",
 "research_question": "if a model's forecast matches the consensus of the other models, is the forecast more reliable?",
 "method": "leave-one-out strict majority, >=3 directional peers",
 "base_field_hit": {
  "n": 4520,
  "hit_rate": 0.4283
 },
 "market_state": {
  "panel": "market_state_6p2",
  "week": {
   "start": "2026-08-31",
   "end": "2026-09-07"
  },
  "prev_week": {
   "start": "2026-08-24",
   "end": "2026-08-31"
  },
  "current": {
   "btc_net_pct": 3.42,
   "realized_vol_ann_pct": 43.35,
   "volume_top5_usd_bn": 16.2391,
   "volume_top5_wow_pct": -17.7,
   "avg_pairwise_corr": 0.78
  },
  "previous_as_published_issue4": {
   "btc_net_pct": -0.07,
   "realized_vol_ann_pct": 30.85,
   "volume_top5_usd_bn": 19.7224,
   "avg_pairwise_corr": 0.85
  },
  "snapshot_line": "BTC net +3.42% (prior -0.07%) - ann vol 43.4% (was 30.9%) - TOP5 volume $16.24B, -18% w/w - pairwise corr 0.78 (was 0.85)",
  "exchange": "binance",
  "audit_filter_applied": true,
  "source": "single-exchange (binance) daily candle recompute, market_state_2026-08-31.json; prior-week values as published in issue #4",
  "audit_note": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #4 after the audit that found the raw table mixes two exchanges; issue #4 values are as published.",
  "prev_week_control": {
   "published_issue3": {
    "btc_net_pct": -0.07,
    "realized_vol_ann_pct": 30.85,
    "volume_top5_usd_bn": 19.7224,
    "avg_pairwise_corr": 0.85
   },
   "published_issue4": {
    "btc_net_pct": -0.07,
    "realized_vol_ann_pct": 30.85,
    "volume_top5_usd_bn": 19.7224,
    "avg_pairwise_corr": 0.85
   },
   "got": {
    "btc_net_pct": -0.07,
    "realized_vol_ann_pct": 30.85,
    "volume_top5_usd_bn": 19.7224,
    "avg_pairwise_corr": 0.85
   },
   "match": {
    "btc_net_pct": true,
    "realized_vol_ann_pct": true,
    "volume_top5_usd_bn": true,
    "avg_pairwise_corr": true
   },
   "tolerance": {
    "btc_net_pct": 0.05,
    "realized_vol_ann_pct": 0.5,
    "volume_top5_usd_bn": 0.05,
    "avg_pairwise_corr": 0.02
   },
   "all_match": true,
   "note": "previous week of this run == published issue-#4 week; the four Snapshot pins must reproduce (key published_issue3 kept for the builder schema, published_issue4 is the add-only alias)"
  },
  "regime_days": {
   "2026-08-31": {
    "move_pct": 1.19,
    "regime": "flat"
   },
   "2026-09-01": {
    "move_pct": -1.5,
    "regime": "trend"
   },
   "2026-09-02": {
    "move_pct": -0.06,
    "regime": "flat"
   },
   "2026-09-03": {
    "move_pct": 4.98,
    "regime": "trend"
   },
   "2026-09-04": {
    "move_pct": -1.97,
    "regime": "trend"
   },
   "2026-09-05": {
    "move_pct": 0.24,
    "regime": "flat"
   },
   "2026-09-06": {
    "move_pct": 0.58,
    "regime": "flat"
   }
  },
  "alive_slice_context": {
   "abs_move_1d_pct": {
    "A": 2.352,
    "B": 2.102
   },
   "btc_realized_vol_ann_pct_hourly": {
    "A": 37.9,
    "B": 35.22
   }
  }
 },
 "tables": {
  "pooled": {
   "agree": {
    "n": 3725,
    "hit_rate": 0.4158,
    "wilson_95": [
     0.4001,
     0.4317
    ]
   },
   "disagree": {
    "n": 26,
    "hit_rate": 0.6923,
    "wilson_95": [
     0.5001,
     0.835
    ]
   },
   "lift_pp": -27.6,
   "thin_or_tied": 769,
   "by_margin": {
    "2": {
     "n": 21,
     "hit_rate": 0.2381
    },
    "3": {
     "n": 456,
     "hit_rate": 0.4474
    },
    "4": {
     "n": 615,
     "hit_rate": 0.5398
    },
    "5": {
     "n": 918,
     "hit_rate": 0.4303
    },
    "6": {
     "n": 1715,
     "hit_rate": 0.3574
    }
   },
   "total_scoreable": 3751,
   "caveat": "Read with context. The disagree bucket is n=26 this week against 36 in issue #4 and 125 in issue #3; the two Wilson intervals do not overlap. This is a single week of dependent observations: not evidence about herding in general, and not a claim that the sign will hold next week."
  },
  "by_cell": [
   {
    "agree": {
     "n": 2402,
     "hit_rate": 0.4467,
     "wilson_95": [
      0.4269,
      0.4667
     ]
    },
    "disagree": {
     "n": 20,
     "hit_rate": 0.65,
     "wilson_95": [
      0.4329,
      0.8188
     ]
    },
    "lift_pp": -20.3,
    "thin_or_tied": 503,
    "by_margin": {
     "2": {
      "n": 9,
      "hit_rate": 0.5556
     },
     "3": {
      "n": 284,
      "hit_rate": 0.4507
     },
     "4": {
      "n": 440,
      "hit_rate": 0.5273
     },
     "5": {
      "n": 570,
      "hit_rate": 0.4719
     },
     "6": {
      "n": 1099,
      "hit_rate": 0.3995
     }
    },
    "n_subjects": 2925,
    "cell": "fh_1h_tf_1h",
    "fh": "1h",
    "tf": "1h"
   },
   {
    "agree": {
     "n": 558,
     "hit_rate": 0.3351,
     "wilson_95": [
      0.2972,
      0.3753
     ]
    },
    "disagree": {
     "n": 3,
     "hit_rate": 0.6667,
     "wilson_95": [
      0.2077,
      0.9385
     ]
    },
    "lift_pp": -33.2,
    "thin_or_tied": 131,
    "by_margin": {
     "2": {
      "n": 3,
      "hit_rate": 0.0
     },
     "3": {
      "n": 84,
      "hit_rate": 0.3214
     },
     "4": {
      "n": 90,
      "hit_rate": 0.5333
     },
     "5": {
      "n": 150,
      "hit_rate": 0.4
     },
     "6": {
      "n": 231,
      "hit_rate": 0.2251
     }
    },
    "n_subjects": 692,
    "cell": "fh_4h_tf_1h",
    "fh": "4h",
    "tf": "1h"
   },
   {
    "agree": {
     "n": 553,
     "hit_rate": 0.3852,
     "wilson_95": [
      0.3455,
      0.4264
     ]
    },
    "disagree": {
     "n": 3,
     "hit_rate": 1.0,
     "wilson_95": [
      0.4385,
      1.0
     ]
    },
    "lift_pp": -61.5,
    "thin_or_tied": 95,
    "by_margin": {
     "2": {
      "n": 9,
      "hit_rate": 0.0
     },
     "3": {
      "n": 48,
      "hit_rate": 0.6875
     },
     "4": {
      "n": 65,
      "hit_rate": 0.4923
     },
     "5": {
      "n": 144,
      "hit_rate": 0.2917
     },
     "6": {
      "n": 287,
      "hit_rate": 0.3693
     }
    },
    "n_subjects": 651,
    "cell": "fh_4h_tf_4h",
    "fh": "4h",
    "tf": "4h"
   },
   {
    "agree": {
     "n": 98,
     "hit_rate": 0.3776,
     "wilson_95": [
      0.2879,
      0.4764
     ]
    },
    "disagree": {
     "n": 0,
     "hit_rate": null,
     "wilson_95": [
      null,
      null
     ]
    },
    "lift_pp": null,
    "thin_or_tied": 30,
    "by_margin": {
     "3": {
      "n": 20,
      "hit_rate": 0.4
     },
     "4": {
      "n": 10,
      "hit_rate": 1.0
     },
     "5": {
      "n": 12,
      "hit_rate": 1.0
     },
     "6": {
      "n": 56,
      "hit_rate": 0.125
     }
    },
    "n_subjects": 128,
    "cell": "fh_1d_tf_1d",
    "fh": "1d",
    "tf": "1d"
   },
   {
    "agree": {
     "n": 114,
     "hit_rate": 0.3421,
     "wilson_95": [
      0.2614,
      0.4331
     ]
    },
    "disagree": {
     "n": 0,
     "hit_rate": null,
     "wilson_95": [
      null,
      null
     ]
    },
    "lift_pp": null,
    "thin_or_tied": 10,
    "by_margin": {
     "3": {
      "n": 20,
      "hit_rate": 0.4
     },
     "4": {
      "n": 10,
      "hit_rate": 1.0
     },
     "5": {
      "n": 42,
      "hit_rate": 0.2857
     },
     "6": {
      "n": 42,
      "hit_rate": 0.2143
     }
    },
    "n_subjects": 124,
    "cell": "fh_1d_tf_4h",
    "fh": "1d",
    "tf": "4h"
   }
  ],
  "by_margin": [
   {
    "margin": 2,
    "n": 21,
    "hit_rate": 0.2381
   },
   {
    "margin": 3,
    "n": 456,
    "hit_rate": 0.4474
   },
   {
    "margin": 4,
    "n": 615,
    "hit_rate": 0.5398
   },
   {
    "margin": 5,
    "n": 918,
    "hit_rate": 0.4303
   },
   {
    "margin": 6,
    "n": 1715,
    "hit_rate": 0.3574
   }
  ],
  "by_margin_week_over_week": [
   {
    "margin": 3,
    "issue4_hit_rate_pct": 49.3,
    "issue5_hit_rate": 0.4474,
    "issue5_n": 456
   },
   {
    "margin": 4,
    "issue4_hit_rate_pct": 47.6,
    "issue5_hit_rate": 0.5398,
    "issue5_n": 615
   },
   {
    "margin": 5,
    "issue4_hit_rate_pct": 35.9,
    "issue5_hit_rate": 0.4303,
    "issue5_n": 918
   },
   {
    "margin": 6,
    "issue4_hit_rate_pct": 39.4,
    "issue5_hit_rate": 0.3574,
    "issue5_n": 1715
   }
  ],
  "per_model": [
   {
    "agree": {
     "n": 557,
     "hit_rate": 0.4255,
     "wilson_95": [
      0.3851,
      0.4669
     ]
    },
    "disagree": {
     "n": 1,
     "hit_rate": 0.0,
     "wilson_95": [
      0.0,
      0.7935
     ]
    },
    "lift_pp": 42.5,
    "thin_or_tied": 70,
    "by_margin": {
     "2": {
      "n": 1,
      "hit_rate": 1.0
     },
     "3": {
      "n": 72,
      "hit_rate": 0.4583
     },
     "4": {
      "n": 93,
      "hit_rate": 0.5699
     },
     "5": {
      "n": 146,
      "hit_rate": 0.4452
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3469
     }
    },
    "model": "claude-fable-5"
   },
   {
    "agree": {
     "n": 489,
     "hit_rate": 0.4233,
     "wilson_95": [
      0.3803,
      0.4675
     ]
    },
    "disagree": {
     "n": 2,
     "hit_rate": 0.5,
     "wilson_95": [
      0.0945,
      0.9055
     ]
    },
    "lift_pp": -7.7,
    "thin_or_tied": 54,
    "by_margin": {
     "2": {
      "n": 1,
      "hit_rate": 1.0
     },
     "3": {
      "n": 52,
      "hit_rate": 0.4615
     },
     "4": {
      "n": 66,
      "hit_rate": 0.5455
     },
     "5": {
      "n": 125,
      "hit_rate": 0.456
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3633
     }
    },
    "model": "claude-opus-5"
   },
   {
    "agree": {
     "n": 560,
     "hit_rate": 0.4268,
     "wilson_95": [
      0.3865,
      0.4681
     ]
    },
    "disagree": {
     "n": 6,
     "hit_rate": 0.8333,
     "wilson_95": [
      0.4365,
      0.9699
     ]
    },
    "lift_pp": -40.7,
    "thin_or_tied": 158,
    "by_margin": {
     "2": {
      "n": 2,
      "hit_rate": 0.0
     },
     "3": {
      "n": 75,
      "hit_rate": 0.44
     },
     "4": {
      "n": 99,
      "hit_rate": 0.5455
     },
     "5": {
      "n": 139,
      "hit_rate": 0.4604
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3592
     }
    },
    "model": "deepseek-v4-pro"
   },
   {
    "agree": {
     "n": 554,
     "hit_rate": 0.4422,
     "wilson_95": [
      0.4014,
      0.4839
     ]
    },
    "disagree": {
     "n": 11,
     "hit_rate": 0.7273,
     "wilson_95": [
      0.4343,
      0.9025
     ]
    },
    "lift_pp": -28.5,
    "thin_or_tied": 196,
    "by_margin": {
     "2": {
      "n": 4,
      "hit_rate": 0.25
     },
     "3": {
      "n": 78,
      "hit_rate": 0.4615
     },
     "4": {
      "n": 98,
      "hit_rate": 0.602
     },
     "5": {
      "n": 129,
      "hit_rate": 0.4419
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3755
     }
    },
    "model": "gemini-3.1-pro"
   },
   {
    "agree": {
     "n": 572,
     "hit_rate": 0.4143,
     "wilson_95": [
      0.3747,
      0.4551
     ]
    },
    "disagree": {
     "n": 0,
     "hit_rate": null,
     "wilson_95": [
      null,
      null
     ]
    },
    "lift_pp": null,
    "thin_or_tied": 117,
    "by_margin": {
     "2": {
      "n": 5,
      "hit_rate": 0.0
     },
     "3": {
      "n": 74,
      "hit_rate": 0.4865
     },
     "4": {
      "n": 99,
      "hit_rate": 0.5051
     },
     "5": {
      "n": 149,
      "hit_rate": 0.4295
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3551
     }
    },
    "model": "gpt-5.6-sol"
   },
   {
    "agree": {
     "n": 428,
     "hit_rate": 0.3598,
     "wilson_95": [
      0.3158,
      0.4063
     ]
    },
    "disagree": {
     "n": 4,
     "hit_rate": 0.5,
     "wilson_95": [
      0.15,
      0.85
     ]
    },
    "lift_pp": -14.0,
    "thin_or_tied": 44,
    "by_margin": {
     "2": {
      "n": 4,
      "hit_rate": 0.25
     },
     "3": {
      "n": 30,
      "hit_rate": 0.3
     },
     "4": {
      "n": 62,
      "hit_rate": 0.4516
     },
     "5": {
      "n": 87,
      "hit_rate": 0.3563
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3469
     }
    },
    "model": "grok-4.6"
   },
   {
    "agree": {
     "n": 565,
     "hit_rate": 0.4071,
     "wilson_95": [
      0.3673,
      0.4481
     ]
    },
    "disagree": {
     "n": 2,
     "hit_rate": 1.0,
     "wilson_95": [
      0.3424,
      1.0
     ]
    },
    "lift_pp": -59.3,
    "thin_or_tied": 130,
    "by_margin": {
     "2": {
      "n": 4,
      "hit_rate": 0.25
     },
     "3": {
      "n": 75,
      "hit_rate": 0.44
     },
     "4": {
      "n": 98,
      "hit_rate": 0.5306
     },
     "5": {
      "n": 143,
      "hit_rate": 0.3986
     },
     "6": {
      "n": 245,
      "hit_rate": 0.3551
     }
    },
    "model": "qwen-3.8-max"
   }
  ],
  "herd_worst_cards": [
   {
    "slot": "2026-09-02T02:01:00+00:00",
    "symbol": "BTC",
    "fh": "1h",
    "tf": "1h",
    "side": "short",
    "n_models": 7,
    "mean_conf": 68.6,
    "hit_rate": 0.0
   },
   {
    "slot": "2026-09-06T12:01:00+00:00",
    "symbol": "SOL",
    "fh": "4h",
    "tf": "1h",
    "side": "long",
    "n_models": 7,
    "mean_conf": 68.4,
    "hit_rate": 0.0
   },
   {
    "slot": "2026-09-06T13:01:00+00:00",
    "symbol": "SOL",
    "fh": "1h",
    "tf": "1h",
    "side": "long",
    "n_models": 7,
    "mean_conf": 68.3,
    "hit_rate": 0.2857
   },
   {
    "slot": "2026-09-03T16:01:00+00:00",
    "symbol": "SOL",
    "fh": "4h",
    "tf": "4h",
    "side": "long",
    "n_models": 7,
    "mean_conf": 68.1,
    "hit_rate": 0.1429
   },
   {
    "slot": "2026-09-02T02:01:00+00:00",
    "symbol": "XRP",
    "fh": "1h",
    "tf": "1h",
    "side": "short",
    "n_models": 7,
    "mean_conf": 67.9,
    "hit_rate": 0.0
   }
  ],
  "counters": {
   "forecasts_total": 10290,
   "ok_in_gate": 9264,
   "out_of_gate": 980,
   "out_of_gate_by_fh": {
    "1w": 490,
    "1m": 490
   },
   "invalid": 46,
   "mature": 9264,
   "mature_directional": 4520,
   "mature_sideways": 4744,
   "pending_next_issue": 0,
   "pending_by_fh": {},
   "late_closes": 0,
   "uptime": [
    {
     "fh": "1h",
     "tf": "1h",
     "slots_seen": 168,
     "slots_expected": 168
    },
    {
     "fh": "4h",
     "tf": "4h",
     "slots_seen": 42,
     "slots_expected": 42
    },
    {
     "fh": "4h",
     "tf": "1h",
     "slots_seen": 42,
     "slots_expected": 42
    },
    {
     "fh": "1d",
     "tf": "1d",
     "slots_seen": 7,
     "slots_expected": 7
    },
    {
     "fh": "1d",
     "tf": "4h",
     "slots_seen": 7,
     "slots_expected": 7
    }
   ],
   "regime_days": {
    "2026-08-31": {
     "move_pct": 1.19,
     "regime": "flat"
    },
    "2026-09-01": {
     "move_pct": -1.5,
     "regime": "trend"
    },
    "2026-09-02": {
     "move_pct": -0.06,
     "regime": "flat"
    },
    "2026-09-03": {
     "move_pct": 4.98,
     "regime": "trend"
    },
    "2026-09-04": {
     "move_pct": -1.97,
     "regime": "trend"
    },
    "2026-09-05": {
     "move_pct": 0.24,
     "regime": "flat"
    },
    "2026-09-06": {
     "move_pct": 0.58,
     "regime": "flat"
    }
   },
   "grok_era": {
    "ids_in_window": [
     "grok-4.6"
    ],
    "n_grok45": 0,
    "n_grok46": 1470,
    "n_grok45_rows": 0,
    "n_grok46_rows": 1470,
    "first_grok46_slot": "2026-08-31T00:01:00+00:00",
    "last_grok45_slot": null,
    "first_grok45_slot": null,
    "flip_at": "2026-08-24T09:22:00+00:00",
    "flip_inside_window": false,
    "prev_window": {
     "n_grok45_rows": 110,
     "n_grok46_rows": 1360
    },
    "merged_model_id": "grok-4.6",
    "merged_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "lineage_ids": [
     "grok-4.5",
     "grok-4.6"
    ],
    "lineage_source": "benchmarks.config.lineage_ids('grok-4.6') + MODEL_SUCCESSION[grok-4.6]=('grok-4.5',)",
    "lineage_warnings": [],
    "rows_remapped_a_plus_b": 110,
    "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- inside the PREVIOUS window (week A); all per-model aggregates use the lineage splice grok-4.5 -> grok-4.6 (one row, labelled 'grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)'). Week B (this window) is pure grok-4.6; week A (prev window) mixes both eras and reproduces issue #4.",
    "cells_with_two_grok_rows": 0
   },
   "series_density": {
    "1d": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": true,
     "n_slots": 7,
     "n_rows": 490,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "1h": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": true,
     "n_slots": 168,
     "n_rows": 5880,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "1m": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": false,
     "n_slots": 7,
     "n_rows": 490,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "1w": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": false,
     "n_slots": 7,
     "n_rows": 490,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "4h": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": true,
     "n_slots": 42,
     "n_rows": 2940,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    }
   },
   "model_roster": {
    "current": [
     "claude-fable-5",
     "claude-opus-5",
     "deepseek-v4-pro",
     "gemini-3.1-pro",
     "gpt-5.6-sol",
     "grok-4.6",
     "qwen-3.8-max"
    ],
    "current_count": 7,
    "current_raw_ids_in_window": [
     "claude-fable-5",
     "claude-opus-5",
     "deepseek-v4-pro",
     "gemini-3.1-pro",
     "gpt-5.6-sol",
     "grok-4.6",
     "qwen-3.8-max"
    ],
    "archived_legacy": [
     "grok-4.5",
     "qwen-3.7-max",
     "claude-opus-4.8"
    ],
    "archived_legacy_count": 3,
    "archived_legacy_detail": [
     {
      "model": "grok-4.5",
      "n_rows_since": 11128,
      "first_slot": "2026-07-15T18:01:00+00:00",
      "last_slot": "2026-08-24T09:01:00+00:00"
     },
     {
      "model": "qwen-3.7-max",
      "n_rows_since": 7479,
      "first_slot": "2026-07-15T18:01:00+00:00",
      "last_slot": "2026-08-05T09:01:00+00:00"
     },
     {
      "model": "claude-opus-4.8",
      "n_rows_since": 6315,
      "first_slot": "2026-07-15T18:01:00+00:00",
      "last_slot": "2026-07-30T08:01:00+00:00"
     }
    ],
    "archived_excluding_lineage_merged": [
     "qwen-3.7-max",
     "claude-opus-4.8"
    ],
    "tracked_total": 10,
    "canon_expected": {
     "current": 7,
     "archived_legacy": 4,
     "tracked_total": 11
    },
    "matches_canon_11_7_4": false,
    "roster_since": "2026-07-11T00:00:00+00:00",
    "note": "current = merged-lineage model ids emitting inside the window (grok-4.5 rows are spliced into grok-4.6); archived legacy = raw ids seen since 2026-07-11 that no longer emit in the window -- grok-4.5 is one of them by raw id even though its rows are spliced. Models retired before 2026-07-11 are not visible to this query."
   }
  }
 },
 "herd_note": "392 cells had 6+ directional models on one side; unanimous_cells_ge6 per weekly metrics",
 "smart_note": "Smart-vs-Outlook deferred: as-of weights pipeline not yet recording; will appear once weights are logged per-slot (methodology 3.2.3 fallback)",
 "key_finding": {
  "label": "KEY FINDING · OBSERVATION (one weekly window)",
  "text": "Herding ran at 99.3% of leave-one-out scoreable calls and the pooled lift came in at -27.6pp: agree-hit 41.6% against 69.2% on disagreement, on a disagree bucket of n=26.",
  "evidence_level": "observation"
 },
 "key_findings": [
  "**OBSERVATION -- 99.3% of the field's leave-one-out calls ran with the herd.** Of 3,751 LOO-scoreable directional calls, 3,725 (99.3%) sided with peers' majority (issue #4: 99.1%); agree-hit was 41.6% [40.0%, 43.2%] vs. 69.2% [50.0%, 83.5%] for the 26 disagreers, a pooled lift of -27.6pp. Descriptive, one week, not a claim of a durable edge.",
  "**The margin curve took a fifth reading.** 3 peers 44.7% (was 49.3%), 4 peers 54.0% (was 47.6%), 5 peers 43.0% (was 35.9%), full unanimity 35.7% (was 39.4%). 3 of the 4 tiers sit above the 42.8% field base, and 3 of the 4 kept the side of the base they took in issue #4.",
  "**Daily-horizon unanimity fell below the field base.** 1d/1d margin-6 hit 12.5% (n=56) and 1d/4h margin-6 hit 21.4% (n=42); issue #4 printed 76.2% and 50.5% on the same two cells.",
  "**Herd failures did not disappear.** 392 cells had 6+ directional models on one side (issue #4: 408); the 5 worst sat at 67.9-68.6 mean confidence and resolved 0.0%-28.6%."
 ],
 "market_check": {
  "label": "MARKET CHECK · OBSERVATION (two adjacent calendar weeks)",
  "text": "OBSERVATION — mean |1d move| 2.35% -> 2.10% (-11% rel), BTC realized vol 37.9% -> 35.2% (ann., hourly); field directional accuracy 42.9% -> 42.8% (-0.1pp), 1 of 7 models improved, sim win-rate up for 1 of 7, field sim PnL -$473.59 -> -$669.03 (adjacent calendar weeks Aug 24-30 vs Aug 31-Sep 6; descriptive, one pair of weeks, not a claim).",
  "robustness": "Robustness: raw price-sign accuracy 44.2% -> 51.6% — the opposite direction to the trade-based hit rule.",
  "weeks": {
   "A": {
    "days": "2026-08-24..2026-08-30",
    "slots": [
     "2026-08-24T00:00:00+00:00",
     "2026-08-31T00:00:00+00:00"
    ],
    "cutoff": "2026-08-31T16:00:00+00:00"
   },
   "B": {
    "days": "2026-08-31..2026-09-06",
    "slots": [
     "2026-08-31T00:00:00+00:00",
     "2026-09-07T00:00:00+00:00"
    ],
    "cutoff": "2026-09-07T16:00:00+00:00"
   }
  },
  "price_source": "crypto_spot/1h@:00",
  "accuracy_field": {
   "A": {
    "n": 4576,
    "hit_rate": 0.4294,
    "wilson_95": [
     0.4151,
     0.4438
    ]
   },
   "B": {
    "n": 4520,
    "hit_rate": 0.4283,
    "wilson_95": [
     0.414,
     0.4428
    ]
   },
   "delta_pp": -0.1
  },
  "accuracy_per_model": [
   {
    "model": "gemini-3.1-pro",
    "model_label": "gemini-3.1-pro",
    "A": {
     "n": 755,
     "hit_rate": 0.4424,
     "wilson_95": [
      0.4073,
      0.478
     ]
    },
    "B": {
     "n": 761,
     "hit_rate": 0.4652,
     "wilson_95": [
      0.43,
      0.5007
     ]
    },
    "delta_pp": 2.3
   },
   {
    "model": "claude-opus-5",
    "model_label": "claude-opus-5",
    "A": {
     "n": 561,
     "hit_rate": 0.4242,
     "wilson_95": [
      0.384,
      0.4655
     ]
    },
    "B": {
     "n": 545,
     "hit_rate": 0.4239,
     "wilson_95": [
      0.383,
      0.4657
     ]
    },
    "delta_pp": -0.0
   },
   {
    "model": "qwen-3.8-max",
    "model_label": "qwen-3.8-max",
    "A": {
     "n": 672,
     "hit_rate": 0.4256,
     "wilson_95": [
      0.3887,
      0.4633
     ]
    },
    "B": {
     "n": 697,
     "hit_rate": 0.4247,
     "wilson_95": [
      0.3885,
      0.4617
     ]
    },
    "delta_pp": -0.1
   },
   {
    "model": "claude-fable-5",
    "model_label": "claude-fable-5",
    "A": {
     "n": 651,
     "hit_rate": 0.4301,
     "wilson_95": [
      0.3926,
      0.4684
     ]
    },
    "B": {
     "n": 628,
     "hit_rate": 0.4283,
     "wilson_95": [
      0.3902,
      0.4674
     ]
    },
    "delta_pp": -0.2
   },
   {
    "model": "deepseek-v4-pro",
    "model_label": "deepseek-v4-pro",
    "A": {
     "n": 729,
     "hit_rate": 0.4458,
     "wilson_95": [
      0.4101,
      0.4821
     ]
    },
    "B": {
     "n": 724,
     "hit_rate": 0.442,
     "wilson_95": [
      0.4062,
      0.4784
     ]
    },
    "delta_pp": -0.4
   },
   {
    "model": "gpt-5.6-sol",
    "model_label": "gpt-5.6-sol",
    "A": {
     "n": 729,
     "hit_rate": 0.4266,
     "wilson_95": [
      0.3912,
      0.4628
     ]
    },
    "B": {
     "n": 689,
     "hit_rate": 0.4165,
     "wilson_95": [
      0.3803,
      0.4537
     ]
    },
    "delta_pp": -1.0
   },
   {
    "model": "grok-4.6",
    "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "A": {
     "n": 479,
     "hit_rate": 0.3987,
     "wilson_95": [
      0.3559,
      0.4432
     ]
    },
    "B": {
     "n": 476,
     "hit_rate": 0.3761,
     "wilson_95": [
      0.3337,
      0.4204
     ]
    },
    "delta_pp": -2.3
   }
  ],
  "raw_sign_field": {
   "A": {
    "n_scored": 4567,
    "hit_rate": 0.4419
   },
   "B": {
    "n_scored": 4507,
    "hit_rate": 0.5156
   }
  },
  "abs_move_1d": {
   "A": 2.352,
   "B": 2.102,
   "delta_pct_rel": -10.6
  },
  "btc_realized_vol_ann_pct_hourly": {
   "A": 37.9,
   "B": 35.22
  },
  "trading_field": {
   "A": {
    "n_trades": 4576,
    "wr": 0.3702,
    "pnl_net_usd": -473.59,
    "pnl_gross_usd": -15.99
   },
   "B": {
    "n_trades": 4520,
    "wr": 0.3564,
    "pnl_net_usd": -669.03,
    "pnl_gross_usd": -217.03
   }
  },
  "verdict": {
   "models_with_hit_improved": "1 of 7",
   "models_with_rawsign_improved": "7 of 7",
   "models_with_wr_improved": "1 of 7",
   "field_hit_delta_pp": -0.1,
   "field_pnl_net": {
    "A": -473.59,
    "B": -669.03
   },
   "abs_move_delta": {
    "1h": {
     "A": 0.401,
     "B": 0.331,
     "delta_pct_rel": -17.5
    },
    "4h": {
     "A": 0.81,
     "B": 0.664,
     "delta_pct_rel": -18.0
    },
    "1d": {
     "A": 2.352,
     "B": 2.102,
     "delta_pct_rel": -10.6
    }
   },
   "btc_vol_ann_pct": {
    "A": 37.9,
    "B": 35.22
   },
   "note": "descriptive, two adjacent calendar weeks; hit rule = methodology v1.1 trade-based; raw_sign = price-sign robustness check (closes at :00 vs slots at :01)"
  },
  "lineage_note": "lineage splice ACTIVE: grok-4.5 rows are aggregated into grok-4.6 (flip 2026-08-24T09:22:00+00:00, inside week A). Week A (2026-08-24..2026-08-30) mixes the 4.5-era (Aug 24 00:00-09:22) with 4.6 and reproduces the published issue #4; week B is pure grok-4.6."
 },
 "why_it_matters": "MarketMania publishes forecasts from 7 frontier model lines into the same slots every week; a natural question for anyone reading the feed is whether a model's call is worth more when it agrees with what everyone else is saying, or whether that agreement is just everyone making the same mistake together. Consensus Watch answers that question empirically, one week at a time, using a leave-one-out design built specifically to avoid the obvious trap of comparing a model to an index that includes its own vote.",
 "how_to_read_this": "For every model M and every forecast it made, we rebuild that slot's consensus using only the OTHER active models in the same symbol / forecast-horizon (FH) / timeframe (TF) cell -- M's own call never counts toward its own consensus. M is scored agree if its side (long or short) matches the strict majority of those peers, and disagree if it sits alone against that majority; a cell needs at least 3 directional peers to count at all, and tied peer splits are excluded rather than forced either way. LOO peers are genuinely external to the model being scored, so the comparison below is not comparing a model to itself. The grok rows are one line, so grok never counts as two peers in a cell. The grok 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits inside the PREVIOUS window (week A, issue #4): every issue-#4 grok number is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\"; this window is pure grok-4.6, so the grok week-over-week row compares a pure-4.6 week with a spliced week.",
 "practical_implications": [
  "Five issues, five margin-curve readings. The tier ordering moved again: 4 peers went 47.6% -> 54.0% while full unanimity went 39.4% -> 35.7%, so a consensus filter still should not carry weight in a config.",
  "The pooled lift is -27.6pp this week against -6.0pp in issue #4 and +17.1pp in issue #3. Three readings, three magnitudes: treat the number as this week's weather, not an edge.",
  "Read the consensus number together with the market check on the same page. The field base moved -0.1pp week over week; 2 of the 4 consensus tiers moved up and 2 down against issue #4 (3 peers -4.6pp, 4 peers +6.4pp, 5 peers +7.1pp, unanimity -3.7pp)."
 ],
 "limitations": [
  "Single week (Aug 31-Sep 6, 2026 UTC). Every finding is descriptive for this window only.",
  "Observations are not independent: the same models watch overlapping symbol / FH / TF cells hour after hour. Wilson intervals here are descriptive, not inferential.",
  "Correlational, not causal: all models see the same market data, so 'agreement' and 'hit' can rise together simply because a slot was easy to read. LOO removes self-match bias, not this confound.",
  "The pooled lift rests on 26 disagreeing calls; per-model counts run 0-11. 769 calls (17.0% of the 4,520-call mature directional universe) were thin or tied and are excluded from every consensus table.",
  "This window ran 3 trend days of 7 (issue #4: 3 of 7); every comparison with issue #4 is a comparison across regimes as well as across weeks. Series density and lineage notes (wave 5): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 31-Sep 6) daily coverage is FULL, as it was in issue #4 (the first full week of the series): the 1w series has slots on 7 of 7 days and the 1M series on 7 of 7 days; both sit outside this report's FH gate (1h/4h/1d) and appear only in the exclusions counter (1w 490, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, inside the PREVIOUS window (week A, issue #4): this window carries 1,470 grok rows and none from the 4.5 era, so every grok number in this issue is a pure grok-4.6 line, while every issue-#4 grok number is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\" -- the grok week-over-week row compares a pure-4.6 week with a spliced week. (3) qwen-3.8-max succeeded qwen-3.7-max on Aug 5, 2026 (fully before this window).",
  "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #4 after the audit that found the raw table mixes two exchanges; issue #4 values are as published.",
  "Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional."
 ],
 "notes": {
  "series_density_and_lineage_wave5": "Series density and lineage notes (wave 5): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 31-Sep 6) daily coverage is FULL, as it was in issue #4 (the first full week of the series): the 1w series has slots on 7 of 7 days and the 1M series on 7 of 7 days; both sit outside this report's FH gate (1h/4h/1d) and appear only in the exclusions counter (1w 490, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, inside the PREVIOUS window (week A, issue #4): this window carries 1,470 grok rows and none from the 4.5 era, so every grok number in this issue is a pure grok-4.6 line, while every issue-#4 grok number is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\" -- the grok week-over-week row compares a pure-4.6 week with a spliced week. (3) qwen-3.8-max succeeded qwen-3.7-max on Aug 5, 2026 (fully before this window).",
  "grok_era": {
   "ids_in_window": [
    "grok-4.6"
   ],
   "n_grok45": 0,
   "n_grok46": 1470,
   "n_grok45_rows": 0,
   "n_grok46_rows": 1470,
   "first_grok46_slot": "2026-08-31T00:01:00+00:00",
   "last_grok45_slot": null,
   "first_grok45_slot": null,
   "flip_at": "2026-08-24T09:22:00+00:00",
   "flip_inside_window": false,
   "prev_window": {
    "n_grok45_rows": 110,
    "n_grok46_rows": 1360
   },
   "merged_model_id": "grok-4.6",
   "merged_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
   "lineage_ids": [
    "grok-4.5",
    "grok-4.6"
   ],
   "lineage_source": "benchmarks.config.lineage_ids('grok-4.6') + MODEL_SUCCESSION[grok-4.6]=('grok-4.5',)",
   "lineage_warnings": [],
   "rows_remapped_a_plus_b": 110,
   "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- inside the PREVIOUS window (week A); all per-model aggregates use the lineage splice grok-4.5 -> grok-4.6 (one row, labelled 'grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)'). Week B (this window) is pure grok-4.6; week A (prev window) mixes both eras and reproduces issue #4.",
   "cells_with_two_grok_rows": 0
  },
  "model_ids_line": [
   "claude-fable-5",
   "claude-opus-5",
   "deepseek-v4-pro",
   "gemini-3.1-pro",
   "gpt-5.6-sol",
   "grok-4.6",
   "qwen-3.8-max"
  ],
  "model_roster": {
   "current": [
    "claude-fable-5",
    "claude-opus-5",
    "deepseek-v4-pro",
    "gemini-3.1-pro",
    "gpt-5.6-sol",
    "grok-4.6",
    "qwen-3.8-max"
   ],
   "current_count": 7,
   "current_raw_ids_in_window": [
    "claude-fable-5",
    "claude-opus-5",
    "deepseek-v4-pro",
    "gemini-3.1-pro",
    "gpt-5.6-sol",
    "grok-4.6",
    "qwen-3.8-max"
   ],
   "archived_legacy": [
    "grok-4.5",
    "qwen-3.7-max",
    "claude-opus-4.8"
   ],
   "archived_legacy_count": 3,
   "archived_legacy_detail": [
    {
     "model": "grok-4.5",
     "n_rows_since": 11128,
     "first_slot": "2026-07-15T18:01:00+00:00",
     "last_slot": "2026-08-24T09:01:00+00:00"
    },
    {
     "model": "qwen-3.7-max",
     "n_rows_since": 7479,
     "first_slot": "2026-07-15T18:01:00+00:00",
     "last_slot": "2026-08-05T09:01:00+00:00"
    },
    {
     "model": "claude-opus-4.8",
     "n_rows_since": 6315,
     "first_slot": "2026-07-15T18:01:00+00:00",
     "last_slot": "2026-07-30T08:01:00+00:00"
    }
   ],
   "archived_excluding_lineage_merged": [
    "qwen-3.7-max",
    "claude-opus-4.8"
   ],
   "tracked_total": 10,
   "canon_expected": {
    "current": 7,
    "archived_legacy": 4,
    "tracked_total": 11
   },
   "matches_canon_11_7_4": false,
   "roster_since": "2026-07-11T00:00:00+00:00",
   "note": "current = merged-lineage model ids emitting inside the window (grok-4.5 rows are spliced into grok-4.6); archived legacy = raw ids seen since 2026-07-11 that no longer emit in the window -- grok-4.5 is one of them by raw id even though its rows are spliced. Models retired before 2026-07-11 are not visible to this query."
  },
  "series_density": {
   "1d": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": true,
    "n_slots": 7,
    "n_rows": 490,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "1h": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": true,
    "n_slots": 168,
    "n_rows": 5880,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "1m": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": false,
    "n_slots": 7,
    "n_rows": 490,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "1w": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": false,
    "n_slots": 7,
    "n_rows": 490,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "4h": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": true,
    "n_slots": 42,
    "n_rows": 2940,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   }
  },
  "week_over_week_note": "Week-over-week columns compare back-to-back windows: Aug 24-30 (issue #4, and week A of this issue's alive slice) vs Aug 31-Sep 6 (this issue). \"Was\" values are the numbers published in issue #4; deltas are computed on unrounded rates.",
  "research_to_date_pin": "Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional.",
  "market_state_audit": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #4 after the audit that found the raw table mixes two exchanges; issue #4 values are as published.",
  "engine": "Engine 1.1 has powered the sandbox since Aug 18, i.e. before this window; no cross-engine PnL comparisons are claimed.",
  "report_template": "PDF built with the canonical mm_report.py template module; preview card with report_preview.render_preview"
 },
 "bibtex_key": "mm_consensus_2026w36",
 "citation": {
  "bibtex_key": "mm_consensus_2026w36",
  "title": "Consensus Watch #5: does agreeing with the crowd make an LLM's market call safer?",
  "author": "MarketMania Research",
  "year": 2026,
  "month": "September",
  "day": 8,
  "url": "https://marketmania.ai/research/reports/consensus-watch-2026-08-31.pdf",
  "note": "Methodology v1.1; window Aug 31-Sep 6, 2026 UTC; source weekly_metrics_2026-08-31.json"
 },
 "living_series_note": "**Issue #5.** Consensus Watch is a living weekly comparison; each issue appends one more week of leave-one-out agreement-vs-hit data. Issue #4 asked whether the pooled agree-vs-disagree sign settles: it reads -27.6pp this week after -6.0pp in issue #4 (agree 41.6% vs disagree 69.2% on n=26), and whether any margin tier repeats its position against the field base two issues running: 3 of the 4 tiers did, and the 4-peer tier came in above the field base. The grok row is pure grok-4.6 this issue; issue #4's grok row was the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\". Engine 1.1 has powered the sandbox since Aug 18, i.e. before this window; no cross-engine PnL comparisons are claimed.",
 "testing_next": [
  "Next issue: does the pooled agree-vs-disagree sign keep the same direction for a third issue running?",
  "Do the same margin tiers hold their side of the field base for a third issue?",
  "Monthly series: the agreement curve across regimes at monthly n."
 ],
 "related_research": [
  {
   "title": "Weekly Calibration #5",
   "url": "https://marketmania.ai/research/reports/weekly-calibration-2026-08-31.pdf"
  },
  {
   "title": "Weekly Model Watch #5",
   "url": "https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf"
  },
  {
   "title": "Consensus Watch #4",
   "url": "https://marketmania.ai/research/reports/consensus-watch-2026-08-24.pdf"
  }
 ],
 "related_research_note": "The three weekly reports publish together as one issue each week; each links straight to the others' PDF and to its own previous issue. Direct links are the posting rule from wave 2 on.",
 "research_to_date": {
  "this_report": {
   "loo_scored_observations": 3751
  },
  "platform": {
   "as_of_cutoff": "2026-09-07",
   "resolved_forecasts": 42905,
   "resolved_forecasts_definition": "directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, at the issue cutoff",
   "resolved_forecasts_since": "2026-07-11",
   "models_tracked": 10,
   "models_tracked_detail": "7 models tracked (current line-up; earlier versions folded into their successors' lineage)",
   "assets": 5,
   "forecast_horizons": 5,
   "cadence": "hourly",
   "published_research_reports": 23
  },
  "source": "weekly_metrics_2026-08-31.json",
  "control_reproduces_wave3_pin": true,
  "pin_note": "Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional."
 },
 "site_entry": {
  "slug": "consensus-watch-2026-08-31",
  "series": "weekly",
  "seriesLabel": "Weekly · Consensus",
  "date": "September 8, 2026",
  "title": "Consensus Watch #5",
  "summary": "9,264 mature forecasts asked whether agreeing with the crowd makes an LLM market call safer. Herding ran at 99.3% of leave-one-out scoreable calls (issue #4: 99.1%), and the pooled lift came in at -27.6pp: agree-hit 41.6% against 69.2% on disagreement, on a disagree bucket of n=26 and a 42.8% field base. The margin curve read 44.7% / 54.0% / 43.0% / 35.7% from 3 peers to full unanimity. Field accuracy moved 42.9% -> 42.8% (-0.1pp) week over week, with 1 of 7 models improving.",
  "stats": [
   "3,751 peer-judged calls",
   "99.3% ran with the herd",
   "lift -27.6pp vs disagree"
  ],
  "manifest_key_finding": "Herding ran at 99.3% of leave-one-out scoreable calls and the pooled lift came in at -27.6pp: agree-hit 41.6% against 69.2% on disagreement (disagree bucket n=26), against a 42.8% field base.",
  "files": {
   "pdf": "/research/reports/consensus-watch-2026-08-31.pdf",
   "md": "/research/reports/consensus-watch-2026-08-31.md",
   "json": "/research/reports/consensus-watch-2026-08-31.json"
  }
 },
 "social": {
  "telegram": {
   "photo_caption_html": "🧭 <b>Consensus Watch #5</b> — weekly series (window Aug 31–Sep 6)\n\nResearch question: if a model's forecast matches the consensus of the other models, is it more reliable?\n\n3,751 leave-one-out scored calls, 7 model lines:\n• Herding ran at <b>99.3%</b> (issue #4: 99.1%) — agree-hit 41.6% vs <b>69.2%</b> when a model left the crowd (field base 42.8%, disagree n=26)\n• Margin curve: 3-peer 44.7%, 4-peer 54.0%, 5-peer 43.0%, unanimity 35.7%\n• Daily horizon: 1d/1d unanimity 12.5%, 1d/4h unanimity 21.4%\n• Market check: field accuracy 42.9% -> 42.8% (-0.1pp), 1 of 7 models improved; avg |1d| move 2.35% -> 2.10%\n\nNote: grok is a pure 4.6 line this window; issue #4's grok row was the 4.5+4.6 splice (flip Aug 24 09:22 UTC)\n\n📄 <a href=\"https://marketmania.ai/research/reports/consensus-watch-2026-08-31.pdf\">Full report (PDF)</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #consensus #research",
   "document_caption_html": "📄 Consensus Watch #5 — full report (PDF).\nWeb copy: <a href=\"https://marketmania.ai/research/reports/consensus-watch-2026-08-31.pdf\">consensus-watch-2026-08-31.pdf</a> · machine-readable: <a href=\"https://marketmania.ai/research/reports/consensus-watch-2026-08-31.json\">JSON</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #consensus #research",
   "caption_len": 951,
   "document_caption_len": 402,
   "limit": 1024
  },
  "x": {
   "main": "Consensus Watch #5 is out.\n\n3,751 peer-judged LLM market calls: 99.3% ran with the herd, and agreeing scored 41.6% against 69.2% for the 26 calls that left it (-27.6pp).\n\nFull report: https://marketmania.ai/research/reports/consensus-watch-2026-08-31.pdf\n\n#AI #Crypto #Research #Consensus",
   "main_tco_len": 241,
   "reply": "Margin curve, w/w (hit):\n3 peers: 49.3 -> 44.7\n4 peers: 47.6 -> 54.0\n5 peers: 35.9 -> 43.0\nunanimity: 39.4 -> 35.7\nField base: 42.9 -> 42.8. LOO 3+ peers, methodology v1.1.\nData: https://marketmania.ai/research/reports/consensus-watch-2026-08-31.json",
   "reply_tco_len": 202,
   "limit_tco": 280,
   "replies_max": 1
  }
 }
}