{
 "slug": "model-watch-2026-09-14",
 "series": "weekly",
 "series_label": "WEEKLY · MODEL WATCH",
 "issue": 7,
 "title": "Weekly Model Watch #7",
 "language": "en",
 "window_start": "2026-09-14T00:00:00+00:00",
 "window_end": "2026-09-21T00:00:00+00:00",
 "window_text": "Sep 14-20, 2026 UTC",
 "cutoff": "2026-09-21T16:00:00+00:00",
 "generated_at": "2026-09-22T07:27:22+00:00",
 "methodology": "v1.1 (2026-08-10)",
 "methodology_hash": "e66c7e8c864a2233",
 "source": "weekly_metrics_2026-09-14.json",
 "hit_rule": "direction: exit tp1/tp2 -> hit, sl -> miss, expiry -> sign of gross pnl",
 "bibtex_key": "mm_modelwatch_2026w38",
 "research_question": "who led the week, which tickers were readable, who saw the turn first, and does a model agreeing with itself across timeframes/horizons make its call stronger?",
 "base_field_hit": {
  "n": 5598,
  "hit_rate": 0.5075
 },
 "market_state": {
  "panel": "market_state_6p2",
  "week": {
   "start": "2026-09-14",
   "end": "2026-09-21"
  },
  "prev_week": {
   "start": "2026-09-07",
   "end": "2026-09-14"
  },
  "current": {
   "btc_net_pct": 5.64,
   "realized_vol_ann_pct": 51.01,
   "volume_top5_usd_bn": 18.7386,
   "volume_top5_wow_pct": 16.8,
   "avg_pairwise_corr": 0.93
  },
  "previous_as_published_issue6": {
   "btc_net_pct": -4.36,
   "realized_vol_ann_pct": 19.68,
   "volume_top5_usd_bn": 16.0492,
   "avg_pairwise_corr": 0.74
  },
  "snapshot_line": "BTC net +5.64% (prior -4.36%) - ann vol 51.0% (was 19.7%) - TOP5 volume $18.74B, +17% w/w - pairwise corr 0.93 (was 0.74)",
  "exchange": "binance",
  "audit_filter_applied": true,
  "source": "single-exchange (binance) daily candle recompute, market_state_2026-09-14.json; prior-week values as published in issue #6",
  "audit_note": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #6 after the audit that found the raw table mixes two exchanges; issue #6 values are as published.",
  "prev_week_control": {
   "published_issue3": {
    "btc_net_pct": -4.36,
    "realized_vol_ann_pct": 19.68,
    "volume_top5_usd_bn": 16.0492,
    "avg_pairwise_corr": 0.74
   },
   "published_issue6": {
    "btc_net_pct": -4.36,
    "realized_vol_ann_pct": 19.68,
    "volume_top5_usd_bn": 16.0492,
    "avg_pairwise_corr": 0.74
   },
   "got": {
    "btc_net_pct": -4.36,
    "realized_vol_ann_pct": 19.68,
    "volume_top5_usd_bn": 16.0492,
    "avg_pairwise_corr": 0.74
   },
   "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-#6 week; the four Snapshot pins must reproduce (key published_issue3 kept for the builder schema, published_issue6 is the add-only alias)"
  },
  "regime_days": {
   "2026-09-14": {
    "move_pct": 1.79,
    "regime": "trend"
   },
   "2026-09-15": {
    "move_pct": -3.35,
    "regime": "trend"
   },
   "2026-09-16": {
    "move_pct": 0.86,
    "regime": "flat"
   },
   "2026-09-17": {
    "move_pct": 0.21,
    "regime": "flat"
   },
   "2026-09-18": {
    "move_pct": 5.82,
    "regime": "trend"
   },
   "2026-09-19": {
    "move_pct": 0.46,
    "regime": "flat"
   },
   "2026-09-20": {
    "move_pct": 0.13,
    "regime": "flat"
   }
  },
  "alive_slice_context": {
   "abs_move_1d_pct": {
    "A": 1.347,
    "B": 2.693
   },
   "btc_realized_vol_ann_pct_hourly": {
    "A": 31.4,
    "B": 35.07
   }
  }
 },
 "key_finding": {
  "label": "KEY FINDING · OBSERVATION (one weekly window)",
  "text": "Field accuracy went 39.6% -> 50.7% (+11.2pp) with 7 of 7 models improving, and no weekly title was awarded, for the seventh issue running.",
  "evidence_level": "observation"
 },
 "tables": {
  "leaderboard": {
   "rows": [
    {
     "model": "gemini-3.1-pro",
     "n": 883,
     "hit_rate": 0.5289,
     "wilson_95": [
      0.4959,
      0.5616
     ],
     "coverage": 0.6639,
     "model_label": "gemini-3.1-pro",
     "insufficient": false
    },
    {
     "model": "claude-fable-5",
     "n": 799,
     "hit_rate": 0.5232,
     "wilson_95": [
      0.4885,
      0.5576
     ],
     "coverage": 0.6008,
     "model_label": "claude-fable-5",
     "insufficient": false
    },
    {
     "model": "claude-opus-5",
     "n": 711,
     "hit_rate": 0.5162,
     "wilson_95": [
      0.4795,
      0.5527
     ],
     "coverage": 0.5346,
     "model_label": "claude-opus-5",
     "insufficient": false
    },
    {
     "model": "qwen-3.8-max",
     "n": 819,
     "hit_rate": 0.5079,
     "wilson_95": [
      0.4737,
      0.5421
     ],
     "coverage": 0.6163,
     "model_label": "qwen-3.8-max",
     "insufficient": false
    },
    {
     "model": "gpt-5.6-sol",
     "n": 821,
     "hit_rate": 0.497,
     "wilson_95": [
      0.4628,
      0.5311
     ],
     "coverage": 0.6173,
     "model_label": "gpt-5.6-sol",
     "insufficient": false
    },
    {
     "model": "grok-4.6",
     "n": 629,
     "hit_rate": 0.496,
     "wilson_95": [
      0.4571,
      0.535
     ],
     "coverage": 0.4733,
     "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
     "insufficient": false
    },
    {
     "model": "deepseek-v4-pro",
     "n": 936,
     "hit_rate": 0.484,
     "wilson_95": [
      0.4521,
      0.516
     ],
     "coverage": 0.7038,
     "model_label": "deepseek-v4-pro",
     "insufficient": false
    }
   ],
   "title_awarded": false,
   "title_rule": ">=5pp gap AND non-overlapping 95% CI, N>=10",
   "top2_gap_pp": 0.6,
   "movement_vs_prior_week": [
    {
     "model": "gemini-3.1-pro",
     "hit_rate_issue6_published": 0.4175,
     "hit_rate_issue7": 0.5289,
     "delta_pp": 11.1,
     "rank_issue7": 1
    },
    {
     "model": "claude-fable-5",
     "hit_rate_issue6_published": 0.3654,
     "hit_rate_issue7": 0.5232,
     "delta_pp": 15.8,
     "rank_issue7": 2
    },
    {
     "model": "claude-opus-5",
     "hit_rate_issue6_published": 0.3647,
     "hit_rate_issue7": 0.5162,
     "delta_pp": 15.1,
     "rank_issue7": 3
    },
    {
     "model": "qwen-3.8-max",
     "hit_rate_issue6_published": 0.4003,
     "hit_rate_issue7": 0.5079,
     "delta_pp": 10.8,
     "rank_issue7": 4
    },
    {
     "model": "gpt-5.6-sol",
     "hit_rate_issue6_published": 0.4006,
     "hit_rate_issue7": 0.497,
     "delta_pp": 9.6,
     "rank_issue7": 5
    },
    {
     "model": "grok-4.6",
     "hit_rate_issue6_published": 0.3958,
     "hit_rate_issue7": 0.496,
     "delta_pp": 10.0,
     "rank_issue7": 6
    },
    {
     "model": "deepseek-v4-pro",
     "hit_rate_issue6_published": 0.414,
     "hit_rate_issue7": 0.484,
     "delta_pp": 7.0,
     "rank_issue7": 7
    }
   ],
   "note": "No weekly title is awarded this issue -- for the seventh time. Title rule: **>=5pp gap AND non-overlapping 95% CI, N>=10** per side. gemini-3.1-pro's lead over claude-fable-5 is 0.6pp and their CIs overlap, so neither condition is met."
  },
  "ticker": {
   "rows": [
    {
     "symbol": "ETH",
     "n": 1116,
     "hit_rate": 0.5717,
     "wilson_95": [
      0.5425,
      0.6004
     ],
     "issue6_hit_rate": 0.2955
    },
    {
     "symbol": "XRP",
     "n": 1108,
     "hit_rate": 0.5036,
     "wilson_95": [
      0.4742,
      0.533
     ],
     "issue6_hit_rate": 0.4216
    },
    {
     "symbol": "BTC",
     "n": 1163,
     "hit_rate": 0.503,
     "wilson_95": [
      0.4743,
      0.5317
     ],
     "issue6_hit_rate": 0.4141
    },
    {
     "symbol": "SOL",
     "n": 1156,
     "hit_rate": 0.4948,
     "wilson_95": [
      0.4661,
      0.5236
     ],
     "issue6_hit_rate": 0.36
    },
    {
     "symbol": "BNB",
     "n": 1055,
     "hit_rate": 0.4626,
     "wilson_95": [
      0.4327,
      0.4927
     ],
     "issue6_hit_rate": 0.4694
    }
   ],
   "best": "ETH",
   "hardest": "BNB",
   "all_above_51pct": false
  },
  "pairs_top": [
   {
    "model": "gemini-3.1-pro",
    "symbol": "ETH",
    "n": 164,
    "hit_rate": 0.5793
   },
   {
    "model": "claude-fable-5",
    "symbol": "ETH",
    "n": 152,
    "hit_rate": 0.5789
   },
   {
    "model": "claude-opus-5",
    "symbol": "ETH",
    "n": 147,
    "hit_rate": 0.5782
   },
   {
    "model": "grok-4.6",
    "symbol": "ETH",
    "n": 135,
    "hit_rate": 0.5778
   },
   {
    "model": "qwen-3.8-max",
    "symbol": "ETH",
    "n": 160,
    "hit_rate": 0.5687
   },
   {
    "model": "gpt-5.6-sol",
    "symbol": "ETH",
    "n": 167,
    "hit_rate": 0.5629
   }
  ],
  "pairs_worst": [
   {
    "model": "gpt-5.6-sol",
    "symbol": "BNB",
    "n": 152,
    "hit_rate": 0.4539
   },
   {
    "model": "deepseek-v4-pro",
    "symbol": "SOL",
    "n": 192,
    "hit_rate": 0.4531
   },
   {
    "model": "grok-4.6",
    "symbol": "SOL",
    "n": 129,
    "hit_rate": 0.4496
   },
   {
    "model": "deepseek-v4-pro",
    "symbol": "BNB",
    "n": 171,
    "hit_rate": 0.4386
   },
   {
    "model": "qwen-3.8-max",
    "symbol": "BNB",
    "n": 149,
    "hit_rate": 0.4362
   }
  ],
  "reversal_events": {
   "definition": "4h grid, trend=sign of prior 24h, counter-move >=2.0% (BTC/ETH) sustained 8h; caller: directional hit, FH 4h/1d, slot in [T-12h, T+FH]",
   "events": [
    {
     "symbol": "ETH",
     "t": "2026-09-20T12:00:00+00:00",
     "new_side": "long",
     "move_8h_pct": 2.22,
     "n_callers": 3,
     "first_caller": {
      "model": "gemini-3.1-pro",
      "slot": "2026-09-20T00:01:00+00:00",
      "fh": "1d",
      "tf": "1d",
      "confidence": 75
     },
     "caller_models": [
      "deepseek-v4-pro",
      "gemini-3.1-pro",
      "gpt-5.6-sol"
     ]
    }
   ],
   "scope_v1": "BTC/ETH only",
   "status": "ok"
  },
  "cross_tf": {
   "fh_4h_tf_4h_vs_1h": {
    "agree": {
     "n": 657,
     "hit_rate": 0.5297,
     "wilson_95": [
      0.4915,
      0.5676
     ],
     "by_regime": {
      "trend": {
       "n": 353,
       "hit_rate": 0.6289
      },
      "flat": {
       "n": 304,
       "hit_rate": 0.4145
      }
     }
    },
    "disagree": {
     "n": 27,
     "hit_rate": 0.4815,
     "wilson_95": [
      0.3074,
      0.6601
     ],
     "by_regime": {
      "trend": {
       "n": 4,
       "hit_rate": 0.75
      },
      "flat": {
       "n": 23,
       "hit_rate": 0.4348
      }
     }
    },
    "mixed_pairs": 214,
    "base": {
     "n": 898,
     "hit_rate": 0.5156
    },
    "lift_pp": 1.4
   },
   "fh_1d_tf_1d_vs_4h": {
    "agree": {
     "n": 133,
     "hit_rate": 0.391,
     "wilson_95": [
      0.3122,
      0.4759
     ],
     "by_regime": {
      "trend": {
       "n": 33,
       "hit_rate": 0.3939
      },
      "flat": {
       "n": 100,
       "hit_rate": 0.39
      }
     }
    },
    "disagree": {
     "n": 3,
     "hit_rate": 1.0,
     "wilson_95": [
      0.4385,
      1.0
     ],
     "by_regime": {
      "trend": {
       "n": 3,
       "hit_rate": 1.0
      },
      "flat": {
       "n": 0,
       "hit_rate": null
      }
     }
    },
    "mixed_pairs": 30,
    "base": {
     "n": 166,
     "hit_rate": 0.3675
    },
    "lift_pp": 2.4
   }
  },
  "cross_fh": {
   "fh_4h_confirmed_by_1h": {
    "agree": {
     "n": 584,
     "hit_rate": 0.5308,
     "wilson_95": [
      0.4903,
      0.571
     ],
     "by_regime": {
      "trend": {
       "n": 323,
       "hit_rate": 0.6223
      },
      "flat": {
       "n": 261,
       "hit_rate": 0.4176
      }
     }
    },
    "disagree": {
     "n": 36,
     "hit_rate": 0.4444,
     "wilson_95": [
      0.2954,
      0.6042
     ],
     "by_regime": {
      "trend": {
       "n": 7,
       "hit_rate": 0.2857
      },
      "flat": {
       "n": 29,
       "hit_rate": 0.4828
      }
     }
    },
    "mixed_pairs": 278,
    "base": {
     "n": 898,
     "hit_rate": 0.5156
    },
    "lift_pp": 1.5
   },
   "fh_1d_confirmed_by_4h": {
    "agree": {
     "n": 104,
     "hit_rate": 0.4327,
     "wilson_95": [
      0.3415,
      0.5286
     ],
     "by_regime": {
      "trend": {
       "n": 23,
       "hit_rate": 0.5652
      },
      "flat": {
       "n": 81,
       "hit_rate": 0.3951
      }
     }
    },
    "disagree": {
     "n": 7,
     "hit_rate": 0.5714,
     "wilson_95": [
      0.2505,
      0.8418
     ],
     "by_regime": {
      "trend": {
       "n": 7,
       "hit_rate": 0.5714
      },
      "flat": {
       "n": 0,
       "hit_rate": null
      }
     }
    },
    "mixed_pairs": 55,
    "base": {
     "n": 166,
     "hit_rate": 0.3675
    },
    "lift_pp": 6.5
   }
  },
  "regime_days": {
   "2026-09-14": {
    "move_pct": 1.79,
    "regime": "trend"
   },
   "2026-09-15": {
    "move_pct": -3.35,
    "regime": "trend"
   },
   "2026-09-16": {
    "move_pct": 0.86,
    "regime": "flat"
   },
   "2026-09-17": {
    "move_pct": 0.21,
    "regime": "flat"
   },
   "2026-09-18": {
    "move_pct": 5.82,
    "regime": "trend"
   },
   "2026-09-19": {
    "move_pct": 0.46,
    "regime": "flat"
   },
   "2026-09-20": {
    "move_pct": 0.13,
    "regime": "flat"
   }
  },
  "side_mix": [
   {
    "model": "claude-fable-5",
    "n": 1330,
    "long": 0.4398,
    "short": 0.1609,
    "sideways": 0.3992,
    "ls_ratio": 2.73
   },
   {
    "model": "claude-opus-5",
    "n": 1330,
    "long": 0.4,
    "short": 0.1346,
    "sideways": 0.4654,
    "ls_ratio": 2.97
   },
   {
    "model": "deepseek-v4-pro",
    "n": 1330,
    "long": 0.4308,
    "short": 0.2729,
    "sideways": 0.2962,
    "ls_ratio": 1.58
   },
   {
    "model": "gemini-3.1-pro",
    "n": 1330,
    "long": 0.4128,
    "short": 0.2511,
    "sideways": 0.3361,
    "ls_ratio": 1.64
   },
   {
    "model": "gpt-5.6-sol",
    "n": 1330,
    "long": 0.3466,
    "short": 0.2707,
    "sideways": 0.3827,
    "ls_ratio": 1.28
   },
   {
    "model": "grok-4.6",
    "n": 1329,
    "long": 0.2626,
    "short": 0.2107,
    "sideways": 0.5267,
    "ls_ratio": 1.25
   },
   {
    "model": "qwen-3.8-max",
    "n": 1329,
    "long": 0.4071,
    "short": 0.2092,
    "sideways": 0.3837,
    "ls_ratio": 1.95
   }
  ],
  "long_short_ratio_field": {
   "value": 1.91,
   "method": "mean of the per-model long-share / short-share ratios"
  },
  "consensus_skew": {
   "cells_ge5_same_side": 685,
   "n_mature": 4356,
   "hit_rate": 0.5053
  },
  "fail_of_week": {
   "model": "gpt-5.6-sol",
   "symbol": "BNB",
   "side": "short",
   "confidence": 84,
   "slot": "2026-09-16T09:01:00+00:00",
   "fh": "1h",
   "tf": "1h",
   "exit_reason": "expiry",
   "pnl_net_usd": -0.67
  },
  "counters": {
   "forecasts_total": 10290,
   "ok_in_gate": 9308,
   "out_of_gate": 973,
   "out_of_gate_by_fh": {
    "1w": 488,
    "1m": 485
   },
   "invalid": 9,
   "mature": 9308,
   "mature_directional": 5598,
   "mature_sideways": 3710,
   "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": {
    "2026-09-14": {
     "move_pct": 1.79,
     "regime": "trend"
    },
    "2026-09-15": {
     "move_pct": -3.35,
     "regime": "trend"
    },
    "2026-09-16": {
     "move_pct": 0.86,
     "regime": "flat"
    },
    "2026-09-17": {
     "move_pct": 0.21,
     "regime": "flat"
    },
    "2026-09-18": {
     "move_pct": 5.82,
     "regime": "trend"
    },
    "2026-09-19": {
     "move_pct": 0.46,
     "regime": "flat"
    },
    "2026-09-20": {
     "move_pct": 0.13,
     "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": "2026-09-14T00:01:00+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": 0,
     "n_grok46_rows": 1470
    },
    "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": 0,
    "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- three windows back (inside issue #4's window); 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)'). Both week A (prev window, issue #6) and week B (this window) are pure grok-4.6.",
    "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": [
      "2026-09-14",
      "2026-09-15",
      "2026-09-16",
      "2026-09-17",
      "2026-09-18",
      "2026-09-19",
      "2026-09-20"
     ],
     "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": [
      "2026-09-14",
      "2026-09-15",
      "2026-09-16",
      "2026-09-17",
      "2026-09-18",
      "2026-09-19",
      "2026-09-20"
     ],
     "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": [
      "2026-09-14",
      "2026-09-15",
      "2026-09-16",
      "2026-09-17",
      "2026-09-18",
      "2026-09-19",
      "2026-09-20"
     ],
     "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": [
      "2026-09-14",
      "2026-09-15",
      "2026-09-16",
      "2026-09-17",
      "2026-09-18",
      "2026-09-19",
      "2026-09-20"
     ],
     "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": [
      "2026-09-14",
      "2026-09-15",
      "2026-09-16",
      "2026-09-17",
      "2026-09-18",
      "2026-09-19",
      "2026-09-20"
     ],
     "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."
   }
  },
  "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
   }
  ],
  "ok_rate_by_model": {
   "claude-fable-5": 1.0,
   "claude-opus-5": 1.0,
   "qwen-3.8-max": 0.9946,
   "grok-4.6": 0.9993,
   "deepseek-v4-pro": 1.0,
   "gemini-3.1-pro": 1.0,
   "gpt-5.6-sol": 1.0
  },
  "calibration_bridge": {
   "best_calibrated_model": "claude-fable-5",
   "brier": 0.2593,
   "hit_rate_leader": "gemini-3.1-pro"
  }
 },
 "key_findings": [
  "**OBSERVATION -- No weekly title again.** gemini-3.1-pro leads descriptively at 52.9% [49.6%, 56.2%]; the gap to #2 (claude-fable-5, 52.3%) is 0.6pp. 7 of 7 models improved week over week; issue #6's leader gemini-3.1-pro is now 1 of 7.",
  "**1 of 5 tickers were readable.** ETH is the week's most readable coin (57.2%, n=1,116) and BNB the hardest at 46.3%; 1 of the 5 assets cleared 51%. Pair of the week: gemini-3.1-pro x ETH 57.9% (n=164); worst: qwen-3.8-max x BNB 43.6% (n=149).",
  "**1 of 1 qualifying turns found a caller.** ETH long +2.22%, 3 caller(s), first gemini-3.1-pro.",
  "**Self-agreement read as a lift in every cell.** Cross-TF/FH lifts ran +1.4pp to +6.5pp (issue #6: -6.0pp to +0.3pp). On trend days, 4h TF-agreement hit 62.9% (n=353) against 41.4% (n=304) on flat days."
 ],
 "market_check": {
  "label": "MARKET CHECK · OBSERVATION (two adjacent calendar weeks)",
  "text": "OBSERVATION — mean |1d move| 1.35% -> 2.69% (+100% rel), BTC realized vol 31.4% -> 35.1% (ann., hourly); field directional accuracy 39.6% -> 50.7% (+11.2pp), 7 of 7 models improved, sim win-rate up for 7 of 7, field sim PnL -$1018.93 -> -$337.72 (adjacent calendar weeks Sep 7-13 vs Sep 14-20; descriptive, one pair of weeks, not a claim).",
  "robustness": "Robustness: raw price-sign accuracy 38.7% -> 54.2% — the same direction as the trade-based hit rule.",
  "weeks": {
   "A": {
    "days": "2026-09-07..2026-09-13",
    "slots": [
     "2026-09-07T00:00:00+00:00",
     "2026-09-14T00:00:00+00:00"
    ],
    "cutoff": "2026-09-14T16:00:00+00:00"
   },
   "B": {
    "days": "2026-09-14..2026-09-20",
    "slots": [
     "2026-09-14T00:00:00+00:00",
     "2026-09-21T00:00:00+00:00"
    ],
    "cutoff": "2026-09-21T16:00:00+00:00"
   }
  },
  "price_source": "crypto_spot/1h@:00",
  "accuracy_field": {
   "A": {
    "n": 4468,
    "hit_rate": 0.3957,
    "wilson_95": [
     0.3815,
     0.4101
    ]
   },
   "B": {
    "n": 5598,
    "hit_rate": 0.5075,
    "wilson_95": [
     0.4944,
     0.5206
    ]
   },
   "delta_pp": 11.2
  },
  "accuracy_per_model": [
   {
    "model": "claude-fable-5",
    "model_label": "claude-fable-5",
    "A": {
     "n": 624,
     "hit_rate": 0.3654,
     "wilson_95": [
      0.3285,
      0.4039
     ]
    },
    "B": {
     "n": 799,
     "hit_rate": 0.5232,
     "wilson_95": [
      0.4885,
      0.5576
     ]
    },
    "delta_pp": 15.8
   },
   {
    "model": "claude-opus-5",
    "model_label": "claude-opus-5",
    "A": {
     "n": 510,
     "hit_rate": 0.3647,
     "wilson_95": [
      0.3241,
      0.4073
     ]
    },
    "B": {
     "n": 711,
     "hit_rate": 0.5162,
     "wilson_95": [
      0.4795,
      0.5527
     ]
    },
    "delta_pp": 15.1
   },
   {
    "model": "gemini-3.1-pro",
    "model_label": "gemini-3.1-pro",
    "A": {
     "n": 673,
     "hit_rate": 0.4175,
     "wilson_95": [
      0.3808,
      0.4552
     ]
    },
    "B": {
     "n": 883,
     "hit_rate": 0.5289,
     "wilson_95": [
      0.4959,
      0.5616
     ]
    },
    "delta_pp": 11.1
   },
   {
    "model": "qwen-3.8-max",
    "model_label": "qwen-3.8-max",
    "A": {
     "n": 707,
     "hit_rate": 0.4003,
     "wilson_95": [
      0.3648,
      0.4368
     ]
    },
    "B": {
     "n": 819,
     "hit_rate": 0.5079,
     "wilson_95": [
      0.4737,
      0.5421
     ]
    },
    "delta_pp": 10.8
   },
   {
    "model": "grok-4.6",
    "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "A": {
     "n": 518,
     "hit_rate": 0.3958,
     "wilson_95": [
      0.3546,
      0.4385
     ]
    },
    "B": {
     "n": 629,
     "hit_rate": 0.496,
     "wilson_95": [
      0.4571,
      0.535
     ]
    },
    "delta_pp": 10.0
   },
   {
    "model": "gpt-5.6-sol",
    "model_label": "gpt-5.6-sol",
    "A": {
     "n": 709,
     "hit_rate": 0.4006,
     "wilson_95": [
      0.3651,
      0.4371
     ]
    },
    "B": {
     "n": 821,
     "hit_rate": 0.497,
     "wilson_95": [
      0.4628,
      0.5311
     ]
    },
    "delta_pp": 9.6
   },
   {
    "model": "deepseek-v4-pro",
    "model_label": "deepseek-v4-pro",
    "A": {
     "n": 727,
     "hit_rate": 0.414,
     "wilson_95": [
      0.3788,
      0.4502
     ]
    },
    "B": {
     "n": 936,
     "hit_rate": 0.484,
     "wilson_95": [
      0.4521,
      0.516
     ]
    },
    "delta_pp": 7.0
   }
  ],
  "raw_sign_field": {
   "A": {
    "n_scored": 4466,
    "hit_rate": 0.3871
   },
   "B": {
    "n_scored": 5583,
    "hit_rate": 0.5416
   }
  },
  "abs_move_1d": {
   "A": 1.347,
   "B": 2.693,
   "delta_pct_rel": 99.9
  },
  "btc_realized_vol_ann_pct_hourly": {
   "A": 31.4,
   "B": 35.07
  },
  "trading_field": {
   "A": {
    "n_trades": 4468,
    "wr": 0.3212,
    "pnl_net_usd": -1018.93,
    "pnl_gross_usd": -572.13
   },
   "B": {
    "n_trades": 5598,
    "wr": 0.4337,
    "pnl_net_usd": -337.72,
    "pnl_gross_usd": 222.08
   }
  },
  "verdict": {
   "models_with_hit_improved": "7 of 7",
   "models_with_rawsign_improved": "7 of 7",
   "models_with_wr_improved": "7 of 7",
   "field_hit_delta_pp": 11.2,
   "field_pnl_net": {
    "A": -1018.93,
    "B": -337.72
   },
   "abs_move_delta": {
    "1h": {
     "A": 0.333,
     "B": 0.369,
     "delta_pct_rel": 10.8
    },
    "4h": {
     "A": 0.605,
     "B": 0.835,
     "delta_pct_rel": 38.0
    },
    "1d": {
     "A": 1.347,
     "B": 2.693,
     "delta_pct_rel": 99.9
    }
   },
   "btc_vol_ann_pct": {
    "A": 31.4,
    "B": 35.07
   },
   "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, three windows back, inside issue #4's window). Week A (2026-09-07..2026-09-13) is pure grok-4.6 (it reproduces the published issue #6); week B is pure grok-4.6."
 },
 "why_it_matters": "MarketMania scores 7 model lines against the same market, hour after hour. Weekly Model Watch asks four practical questions about that week of calls: who led the week, which tickers were readable, who saw the turn first, and does a model agreeing with itself across timeframes/horizons make its call stronger? Each question gets its own block below, with its own table, its own N, and its own honesty caveats. The grok row is a pure grok-4.6 week in this issue and was one in issue #6 as well -- the 4.5 -> 4.6 flip sits three windows back -- so read the grok week-over-week row as a line, not a single model build.",
 "how_to_read_this": "Every block below is scored on the same pool: matured directional calls inside the 1h/4h/1d FH gate, hit by the methodology v1.1 direction rule. The weekly title is awarded only on a >=5pp gap AND non-overlapping 95% Wilson CIs with N>=10 per side -- a descriptive lead is not a title. Reversals use the 4h grid with a >=2.0% counter-move sustained 8h, BTC/ETH only in v1. Cross-TF/FH blocks ask whether a model agreeing with itself does any better, and are split trend vs flat wherever the week supplies both sides. The grok 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits three windows back (issue #4): both weeks of this issue are pure grok-4.6, with no 4.5-era rows in either window, so the grok week-over-week row is the second pure-4.6 against pure-4.6 comparison of the series.",
 "practical_implications": [
  "Do not read this week's leaderboard as a skill ranking: no title was awarded and the top CIs overlap.",
  "Treat per-ticker readability as weekly weather. Seven issues, and the best ticker changed hands in every one except issue #6 (SOL -> BNB -> BTC -> SOL -> BNB -> BNB -> ETH): ETH leads this week at 57.2%, after two issues of BNB.",
  "Self-agreement lifts run +1.4pp to +6.5pp with small disagree cells; carry the regime split into the monthly test, do not act on it."
 ],
 "limitations": [
  "Single week (Sep 14-20, 2026 UTC). Descriptive for this window only; no claim about next week or any model's underlying skill. This window ran 3 trend days of 7 (issue #6: 2 of 7); every comparison with issue #6 is a comparison across regimes as well as across weeks.",
  "Observations inside one window are not independent, and the 95% Wilson CIs shown throughout are descriptive, not inferential.",
  "The reversal detector remains v1 and BTC/ETH-only, on price series reconstructed from trade entry prices. 13 events across seven issues cannot characterise detector performance either way.",
  "Cross-TF and cross-FH disagreement cells are small (n=27, 3, 36, 7); cells under N=10 are insufficient and never used to rank models. Week-over-week deltas are computed on unrounded rates and may differ by 0.1pp from the difference of the rounded columns.",
  "Model lines in this 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. DeepSeek announced the retirement of V4 Pro in favour of V4.1 Flash from Sep 14 (Beijing noon); as of Sep 22 the API still returns deepseek-v4-pro as the served model for our requests and the per-answer latency, token and cost series of Sep 1-22 show no regime change, so this issue keeps the label; it will change the day the served id changes. Series density and lineage notes (wave 7): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Sep 14-20) daily coverage is FULL, as it was in issues #4, #5 and #6: 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 488, 1M 485). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, three windows back (inside the issue-#4 window): this window carries 1,470 grok rows and none from the 4.5 era, and neither does week A (the issue-#6 window), so every grok number in this issue and in issue #6 is a pure grok-4.6 line -- the grok week-over-week row is the second pure-4.6 against pure-4.6 comparison of the series. (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 #6 after the audit that found the raw table mixes two exchanges; issue #6 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 21 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4, #5 and #6, which printed 47,373 at the Sep 14 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 52,971 is an additive step under one definition; counter deltas against issue #5 and earlier remain definitional."
 ],
 "notes": {
  "series_density_and_lineage_wave7": "Series density and lineage notes (wave 7): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Sep 14-20) daily coverage is FULL, as it was in issues #4, #5 and #6: 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 488, 1M 485). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, three windows back (inside the issue-#4 window): this window carries 1,470 grok rows and none from the 4.5 era, and neither does week A (the issue-#6 window), so every grok number in this issue and in issue #6 is a pure grok-4.6 line -- the grok week-over-week row is the second pure-4.6 against pure-4.6 comparison of the series. (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": "2026-09-14T00:01:00+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": 0,
    "n_grok46_rows": 1470
   },
   "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": 0,
   "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- three windows back (inside issue #4's window); 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)'). Both week A (prev window, issue #6) and week B (this window) are pure grok-4.6.",
   "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": [
     "2026-09-14",
     "2026-09-15",
     "2026-09-16",
     "2026-09-17",
     "2026-09-18",
     "2026-09-19",
     "2026-09-20"
    ],
    "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": [
     "2026-09-14",
     "2026-09-15",
     "2026-09-16",
     "2026-09-17",
     "2026-09-18",
     "2026-09-19",
     "2026-09-20"
    ],
    "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": [
     "2026-09-14",
     "2026-09-15",
     "2026-09-16",
     "2026-09-17",
     "2026-09-18",
     "2026-09-19",
     "2026-09-20"
    ],
    "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": [
     "2026-09-14",
     "2026-09-15",
     "2026-09-16",
     "2026-09-17",
     "2026-09-18",
     "2026-09-19",
     "2026-09-20"
    ],
    "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": [
     "2026-09-14",
     "2026-09-15",
     "2026-09-16",
     "2026-09-17",
     "2026-09-18",
     "2026-09-19",
     "2026-09-20"
    ],
    "missing_days": []
   }
  },
  "week_over_week_note": "Week-over-week columns compare back-to-back windows: Sep 7-13 (issue #6, and week A of this issue's alive slice) vs Sep 14-20 (this issue). \"Was\" values are the numbers published in issue #6; 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 21 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4, #5 and #6, which printed 47,373 at the Sep 14 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 52,971 is an additive step under one definition; counter deltas against issue #5 and earlier remain definitional.",
  "market_state_audit": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #6 after the audit that found the raw table mixes two exchanges; issue #6 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"
 },
 "living_series_note": "**Issue #7.** Weekly Model Watch is a living series. Issue #6's open questions -- does gemini-3.1-pro hold the top rank for a third issue and does the title rule ever fire, and does a qualifying reversal ever fire outside ETH or a second turn in the same pair draw a caller? -- read this issue as: gemini-3.1-pro at the top (52.9%), the line that also led issue #6 (gemini-3.1-pro), so the rank held for a third issue; no weekly title was awarded, for the seventh issue running, and 1 of 1 qualifying turns found a caller, on turns in ETH only. The grok row is pure grok-4.6 this issue and was pure grok-4.6 in issue #6; the 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits three windows back. 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 gemini-3.1-pro hold the top rank for a fourth issue, and does the title rule ever fire?",
  "Does a qualifying reversal ever fire outside ETH, and does a second turn in the same pair ever draw a caller?",
  "Monthly series: cross-TF/FH lifts with regime control across mixed weeks (trend vs. flat), block bootstrap."
 ],
 "related_research": [
  {
   "title": "Consensus Watch #7",
   "url": "https://marketmania.ai/research/reports/consensus-watch-2026-09-14.pdf"
  },
  {
   "title": "Weekly Calibration #7",
   "url": "https://marketmania.ai/research/reports/weekly-calibration-2026-09-14.pdf"
  },
  {
   "title": "Weekly Model Watch #6",
   "url": "https://marketmania.ai/research/reports/model-watch-2026-09-07.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": {
   "scored_observations": 5598
  },
  "platform": {
   "as_of_cutoff": "2026-09-21",
   "resolved_forecasts": 52971,
   "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": 33
  },
  "source": "weekly_metrics_2026-09-14.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 21 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4, #5 and #6, which printed 47,373 at the Sep 14 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 52,971 is an additive step under one definition; counter deltas against issue #5 and earlier remain definitional."
 },
 "citation": {
  "bibtex_key": "mm_modelwatch_2026w38",
  "title": "Weekly Model Watch #7: weekly leaderboard, ticker/pair reads and cross-confirmation, Sep 14-20 2026",
  "author": "MarketMania Research",
  "year": 2026,
  "month": "September",
  "day": 22,
  "url": "https://marketmania.ai/research/reports/model-watch-2026-09-14.pdf",
  "note": "Methodology v1.1; window Sep 14-20, 2026 UTC; source weekly_metrics_2026-09-14.json"
 },
 "site_entry": {
  "slug": "model-watch-2026-09-14",
  "series": "weekly",
  "seriesLabel": "Weekly · Model Watch",
  "date": "September 22, 2026",
  "title": "Weekly Model Watch #7",
  "summary": "5,598 directional calls, four questions: who led the week, which tickers were readable, who saw the turn first, does self-agreement help. No weekly title again — gemini-3.1-pro led at 52.9% (+11.1pp w/w) against claude-fable-5 at 52.3%, a gap of 0.6pp. ETH was the most readable ticker at 57.2% and BNB the hardest at 46.3%, with 1 of 5 assets above 51 percent. Reversals: 1 of 1 qualifying turns found a caller. Field accuracy moved 39.6% -> 50.7% (+11.2pp) with 7 of 7 models improving.",
  "stats": [
   "5,598 directional calls",
   "No weekly title again",
   "field +11.2pp w/w"
  ],
  "manifest_key_finding": "Field accuracy moved 39.6% -> 50.7% (+11.2pp) with 7 of 7 models improving; gemini-3.1-pro led the week at 52.9% (+11.1pp w/w) and no weekly title was awarded, for the seventh issue running.",
  "files": {
   "pdf": "/research/reports/model-watch-2026-09-14.pdf",
   "md": "/research/reports/model-watch-2026-09-14.md",
   "json": "/research/reports/model-watch-2026-09-14.json"
  }
 },
 "social": {
  "telegram": {
   "photo_caption_html": "📊 <b>Weekly Model Watch #7</b> — weekly series (window Sep 14–20)\n\nWho led, what was readable, who saw the turn, does self-agreement help? 5,598 directional calls, 7 model lines:\n• <b>No weekly title again</b> — gemini-3.1-pro leads at <b>52.9%</b> (+11.1pp w/w) against claude-fable-5 at 52.3%, a 0.6pp gap\n• Tickers: <b>ETH</b> most readable (57.2%); BNB hardest at 46.3% — 1 of 5 assets above 51%\n• Reversals: 1 of 1 qualifying turns found a caller\n• Market check: field accuracy 39.6% -> 50.7% (+11.2pp), 7 of 7 improved; |1d| move 1.35% -> 2.69%\n\nNote: grok is a pure 4.6 line in both weeks of this issue; the 4.5 -> 4.6 flip (Aug 24 09:22 UTC) sits three windows back\n\n📄 <a href=\"https://marketmania.ai/research/reports/model-watch-2026-09-14.pdf\">Full report (PDF)</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #models #research",
   "document_caption_html": "📄 Weekly Model Watch #7 — full report (PDF).\nWeb copy: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-09-14.pdf\">model-watch-2026-09-14.pdf</a> · machine-readable: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-09-14.json\">JSON</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #models #research",
   "caption_len": 896,
   "document_caption_len": 390,
   "limit": 1024
  },
  "x": {
   "main": "Weekly Model Watch #7 is out.\n\n5,598 directional LLM calls: field accuracy 39.6% -> 50.7% w/w, 7 of 7 models improved. gemini-3.1-pro led at 52.9%; ETH was the most readable ticker.\n\nFull report: https://marketmania.ai/research/reports/model-watch-2026-09-14.pdf\n\n#AI #Crypto #Research #LLM",
   "main_tco_len": 247,
   "reply": "Hit-rate, w/w:\ngemini-3.1: 41.8 -> 52.9\nfable-5: 36.5 -> 52.3\nopus-5: 36.5 -> 51.6\nqwen-3.8: 40.0 -> 50.8\ngpt-5.6: 40.1 -> 49.7\ngrok-4.6: 39.6 -> 49.6\ndeepseek: 41.4 -> 48.4\nField: 39.6 -> 50.7. Wilson CIs; methodology v1.1.\nData: https://marketmania.ai/research/reports/model-watch-2026-09-14.json",
   "reply_tco_len": 254,
   "limit_tco": 280,
   "replies_max": 1
  }
 }
}