{
 "slug": "model-watch-2026-09-21",
 "series": "weekly",
 "series_label": "WEEKLY · MODEL WATCH",
 "issue": 8,
 "title": "Weekly Model Watch #8",
 "language": "en",
 "window_start": "2026-09-21T00:00:00+00:00",
 "window_end": "2026-09-28T00:00:00+00:00",
 "window_text": "Sep 21-27, 2026 UTC",
 "cutoff": "2026-09-28T16:00:00+00:00",
 "generated_at": "2026-09-30T05:45:53+00:00",
 "methodology": "v1.1 (2026-08-10)",
 "methodology_hash": "e66c7e8c864a2233",
 "source": "weekly_metrics_2026-09-21.json",
 "hit_rule": "direction: exit tp1/tp2 -> hit, sl -> miss, expiry -> sign of gross pnl",
 "bibtex_key": "mm_modelwatch_2026w39",
 "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": 4958,
  "hit_rate": 0.4683
 },
 "market_state": {
  "panel": "market_state_6p2",
  "week": {
   "start": "2026-09-21",
   "end": "2026-09-28"
  },
  "prev_week": {
   "start": "2026-09-14",
   "end": "2026-09-21"
  },
  "current": {
   "btc_net_pct": 4.06,
   "realized_vol_ann_pct": 52.25,
   "volume_top5_usd_bn": 22.4265,
   "volume_top5_wow_pct": 19.7,
   "avg_pairwise_corr": 0.71
  },
  "previous_as_published_issue7": {
   "btc_net_pct": 5.64,
   "realized_vol_ann_pct": 51.01,
   "volume_top5_usd_bn": 18.7386,
   "avg_pairwise_corr": 0.93
  },
  "snapshot_line": "BTC net +4.06% (prior +5.64%) - ann vol 52.2% (was 51.0%) - TOP5 volume $22.43B, +20% w/w - pairwise corr 0.71 (was 0.93)",
  "exchange": "binance",
  "audit_filter_applied": true,
  "source": "single-exchange (binance) daily candle recompute, market_state_2026-09-21.json; prior-week values as published in issue #7",
  "audit_note": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #7 after the audit that found the raw table mixes two exchanges; issue #7 values are as published.",
  "prev_week_control": {
   "published_issue3": {
    "btc_net_pct": 5.64,
    "realized_vol_ann_pct": 51.01,
    "volume_top5_usd_bn": 18.7386,
    "avg_pairwise_corr": 0.93
   },
   "published_issue7": {
    "btc_net_pct": 5.64,
    "realized_vol_ann_pct": 51.01,
    "volume_top5_usd_bn": 18.7386,
    "avg_pairwise_corr": 0.93
   },
   "got": {
    "btc_net_pct": 5.64,
    "realized_vol_ann_pct": 51.01,
    "volume_top5_usd_bn": 18.7386,
    "avg_pairwise_corr": 0.93
   },
   "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-#7 week; the four Snapshot pins must reproduce (key published_issue3 kept for the builder schema, published_issue7 is the add-only alias)"
  },
  "regime_days": {
   "2026-09-21": {
    "move_pct": 6.39,
    "regime": "trend"
   },
   "2026-09-22": {
    "move_pct": -0.4,
    "regime": "flat"
   },
   "2026-09-23": {
    "move_pct": -2.04,
    "regime": "trend"
   },
   "2026-09-24": {
    "move_pct": 0.02,
    "regime": "flat"
   },
   "2026-09-25": {
    "move_pct": -0.46,
    "regime": "flat"
   },
   "2026-09-26": {
    "move_pct": 0.36,
    "regime": "flat"
   },
   "2026-09-27": {
    "move_pct": 0.2,
    "regime": "flat"
   }
  },
  "alive_slice_context": {
   "abs_move_1d_pct": {
    "A": 2.693,
    "B": 1.835
   },
   "btc_realized_vol_ann_pct_hourly": {
    "A": 35.07,
    "B": 35.8
   }
  }
 },
 "key_finding": {
  "label": "KEY FINDING · OBSERVATION (one weekly window)",
  "text": "Field accuracy went 50.7% -> 46.8% (-3.9pp) with 0 of 7 models improving, and no weekly title was awarded, for the eighth issue running.",
  "evidence_level": "observation"
 },
 "tables": {
  "leaderboard": {
   "rows": [
    {
     "model": "gemini-3.1-pro",
     "n": 794,
     "hit_rate": 0.5139,
     "wilson_95": [
      0.4791,
      0.5485
     ],
     "coverage": 0.597,
     "model_label": "gemini-3.1-pro",
     "insufficient": false
    },
    {
     "model": "claude-opus-5",
     "n": 620,
     "hit_rate": 0.4758,
     "wilson_95": [
      0.4368,
      0.5151
     ],
     "coverage": 0.4662,
     "model_label": "claude-opus-5",
     "insufficient": false
    },
    {
     "model": "grok-4.6",
     "n": 537,
     "hit_rate": 0.4693,
     "wilson_95": [
      0.4274,
      0.5116
     ],
     "coverage": 0.4059,
     "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
     "insufficient": false
    },
    {
     "model": "gpt-5.6-sol",
     "n": 744,
     "hit_rate": 0.4677,
     "wilson_95": [
      0.4321,
      0.5037
     ],
     "coverage": 0.5594,
     "model_label": "gpt-5.6-sol",
     "insufficient": false
    },
    {
     "model": "claude-fable-5",
     "n": 736,
     "hit_rate": 0.4674,
     "wilson_95": [
      0.4316,
      0.5035
     ],
     "coverage": 0.5534,
     "model_label": "claude-fable-5",
     "insufficient": false
    },
    {
     "model": "qwen-3.8-max",
     "n": 750,
     "hit_rate": 0.4453,
     "wilson_95": [
      0.4101,
      0.4811
     ],
     "coverage": 0.5682,
     "model_label": "qwen-3.8-max",
     "insufficient": false
    },
    {
     "model": "deepseek-v4-pro",
     "n": 777,
     "hit_rate": 0.4389,
     "wilson_95": [
      0.4044,
      0.474
     ],
     "coverage": 0.5931,
     "model_label": "deepseek-v4-pro",
     "insufficient": false
    }
   ],
   "title_awarded": false,
   "title_rule": ">=5pp gap AND non-overlapping 95% CI, N>=10",
   "top2_gap_pp": 3.8,
   "movement_vs_prior_week": [
    {
     "model": "gemini-3.1-pro",
     "hit_rate_issue7_published": 0.5289,
     "hit_rate_issue8": 0.5139,
     "delta_pp": -1.5,
     "rank_issue8": 1
    },
    {
     "model": "claude-opus-5",
     "hit_rate_issue7_published": 0.5162,
     "hit_rate_issue8": 0.4758,
     "delta_pp": -4.0,
     "rank_issue8": 2
    },
    {
     "model": "grok-4.6",
     "hit_rate_issue7_published": 0.496,
     "hit_rate_issue8": 0.4693,
     "delta_pp": -2.7,
     "rank_issue8": 3
    },
    {
     "model": "gpt-5.6-sol",
     "hit_rate_issue7_published": 0.497,
     "hit_rate_issue8": 0.4677,
     "delta_pp": -2.9,
     "rank_issue8": 4
    },
    {
     "model": "claude-fable-5",
     "hit_rate_issue7_published": 0.5232,
     "hit_rate_issue8": 0.4674,
     "delta_pp": -5.6,
     "rank_issue8": 5
    },
    {
     "model": "qwen-3.8-max",
     "hit_rate_issue7_published": 0.5079,
     "hit_rate_issue8": 0.4453,
     "delta_pp": -6.3,
     "rank_issue8": 6
    },
    {
     "model": "deepseek-v4-pro",
     "hit_rate_issue7_published": 0.484,
     "hit_rate_issue8": 0.4389,
     "delta_pp": -4.5,
     "rank_issue8": 7
    }
   ],
   "note": "No weekly title is awarded this issue -- for the eighth time. Title rule: **>=5pp gap AND non-overlapping 95% CI, N>=10** per side. gemini-3.1-pro's lead over claude-opus-5 is 3.8pp and their CIs overlap, so neither condition is met."
  },
  "ticker": {
   "rows": [
    {
     "symbol": "XRP",
     "n": 1165,
     "hit_rate": 0.4996,
     "wilson_95": [
      0.4709,
      0.5282
     ],
     "issue7_hit_rate": 0.5036
    },
    {
     "symbol": "BTC",
     "n": 919,
     "hit_rate": 0.4886,
     "wilson_95": [
      0.4564,
      0.5209
     ],
     "issue7_hit_rate": 0.503
    },
    {
     "symbol": "SOL",
     "n": 1190,
     "hit_rate": 0.4874,
     "wilson_95": [
      0.4591,
      0.5158
     ],
     "issue7_hit_rate": 0.4948
    },
    {
     "symbol": "ETH",
     "n": 891,
     "hit_rate": 0.4579,
     "wilson_95": [
      0.4254,
      0.4907
     ],
     "issue7_hit_rate": 0.5717
    },
    {
     "symbol": "BNB",
     "n": 793,
     "hit_rate": 0.3821,
     "wilson_95": [
      0.3489,
      0.4164
     ],
     "issue7_hit_rate": 0.4626
    }
   ],
   "best": "XRP",
   "hardest": "BNB",
   "all_above_51pct": false
  },
  "pairs_top": [
   {
    "model": "gemini-3.1-pro",
    "symbol": "SOL",
    "n": 186,
    "hit_rate": 0.543
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "BTC",
    "n": 154,
    "hit_rate": 0.5325
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "XRP",
    "n": 181,
    "hit_rate": 0.5304
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "ETH",
    "n": 142,
    "hit_rate": 0.5211
   },
   {
    "model": "gpt-5.6-sol",
    "symbol": "XRP",
    "n": 177,
    "hit_rate": 0.5198
   },
   {
    "model": "grok-4.6",
    "symbol": "XRP",
    "n": 136,
    "hit_rate": 0.5147
   }
  ],
  "pairs_worst": [
   {
    "model": "grok-4.6",
    "symbol": "BNB",
    "n": 86,
    "hit_rate": 0.3837
   },
   {
    "model": "deepseek-v4-pro",
    "symbol": "BNB",
    "n": 124,
    "hit_rate": 0.371
   },
   {
    "model": "qwen-3.8-max",
    "symbol": "BNB",
    "n": 130,
    "hit_rate": 0.3692
   },
   {
    "model": "claude-opus-5",
    "symbol": "BNB",
    "n": 86,
    "hit_rate": 0.3605
   },
   {
    "model": "claude-fable-5",
    "symbol": "BNB",
    "n": 114,
    "hit_rate": 0.3596
   }
  ],
  "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": "BTC",
     "t": "2026-09-23T08:00:00+00:00",
     "new_side": "short",
     "move_8h_pct": -2.9,
     "n_callers": 1,
     "first_caller": {
      "model": "gemini-3.1-pro",
      "slot": "2026-09-24T00:01:00+00:00",
      "fh": "1d",
      "tf": "4h",
      "confidence": 65
     },
     "caller_models": [
      "gemini-3.1-pro"
     ]
    },
    {
     "symbol": "ETH",
     "t": "2026-09-23T08:00:00+00:00",
     "new_side": "short",
     "move_8h_pct": -3.73,
     "n_callers": 6,
     "first_caller": {
      "model": "gpt-5.6-sol",
      "slot": "2026-09-23T08:01:00+00:00",
      "fh": "4h",
      "tf": "1h",
      "confidence": 67
     },
     "caller_models": [
      "deepseek-v4-pro",
      "gemini-3.1-pro",
      "gpt-5.6-sol",
      "grok-4.6",
      "qwen-3.8-max"
     ]
    }
   ],
   "scope_v1": "BTC/ETH only",
   "status": "ok"
  },
  "cross_tf": {
   "fh_4h_tf_4h_vs_1h": {
    "agree": {
     "n": 578,
     "hit_rate": 0.4429,
     "wilson_95": [
      0.4029,
      0.4836
     ],
     "by_regime": {
      "trend": {
       "n": 264,
       "hit_rate": 0.6402
      },
      "flat": {
       "n": 314,
       "hit_rate": 0.2771
      }
     }
    },
    "disagree": {
     "n": 21,
     "hit_rate": 0.4286,
     "wilson_95": [
      0.2447,
      0.6345
     ],
     "by_regime": {
      "trend": {
       "n": 4,
       "hit_rate": 0.25
      },
      "flat": {
       "n": 17,
       "hit_rate": 0.4706
      }
     }
    },
    "mixed_pairs": 173,
    "base": {
     "n": 773,
     "hit_rate": 0.4683
    },
    "lift_pp": -2.5
   },
   "fh_1d_tf_1d_vs_4h": {
    "agree": {
     "n": 135,
     "hit_rate": 0.4889,
     "wilson_95": [
      0.406,
      0.5723
     ],
     "by_regime": {
      "trend": {
       "n": 64,
       "hit_rate": 0.6094
      },
      "flat": {
       "n": 71,
       "hit_rate": 0.3803
      }
     }
    },
    "disagree": {
     "n": 8,
     "hit_rate": 0.5,
     "wilson_95": [
      0.2152,
      0.7848
     ],
     "by_regime": {
      "trend": {
       "n": 0,
       "hit_rate": null
      },
      "flat": {
       "n": 8,
       "hit_rate": 0.5
      }
     }
    },
    "mixed_pairs": 63,
    "base": {
     "n": 206,
     "hit_rate": 0.4806
    },
    "lift_pp": 0.8
   }
  },
  "cross_fh": {
   "fh_4h_confirmed_by_1h": {
    "agree": {
     "n": 513,
     "hit_rate": 0.4581,
     "wilson_95": [
      0.4154,
      0.5014
     ],
     "by_regime": {
      "trend": {
       "n": 248,
       "hit_rate": 0.6492
      },
      "flat": {
       "n": 265,
       "hit_rate": 0.2792
      }
     }
    },
    "disagree": {
     "n": 25,
     "hit_rate": 0.4,
     "wilson_95": [
      0.234,
      0.5926
     ],
     "by_regime": {
      "trend": {
       "n": 7,
       "hit_rate": 0.4286
      },
      "flat": {
       "n": 18,
       "hit_rate": 0.3889
      }
     }
    },
    "mixed_pairs": 234,
    "base": {
     "n": 773,
     "hit_rate": 0.4683
    },
    "lift_pp": -1.0
   },
   "fh_1d_confirmed_by_4h": {
    "agree": {
     "n": 106,
     "hit_rate": 0.5189,
     "wilson_95": [
      0.4248,
      0.6117
     ],
     "by_regime": {
      "trend": {
       "n": 48,
       "hit_rate": 0.7292
      },
      "flat": {
       "n": 58,
       "hit_rate": 0.3448
      }
     }
    },
    "disagree": {
     "n": 9,
     "hit_rate": 0.5556,
     "wilson_95": [
      0.2666,
      0.8112
     ],
     "by_regime": {
      "trend": {
       "n": 0,
       "hit_rate": null
      },
      "flat": {
       "n": 9,
       "hit_rate": 0.5556
      }
     }
    },
    "mixed_pairs": 91,
    "base": {
     "n": 206,
     "hit_rate": 0.4806
    },
    "lift_pp": 3.8
   }
  },
  "regime_days": {
   "2026-09-21": {
    "move_pct": 6.39,
    "regime": "trend"
   },
   "2026-09-22": {
    "move_pct": -0.4,
    "regime": "flat"
   },
   "2026-09-23": {
    "move_pct": -2.04,
    "regime": "trend"
   },
   "2026-09-24": {
    "move_pct": 0.02,
    "regime": "flat"
   },
   "2026-09-25": {
    "move_pct": -0.46,
    "regime": "flat"
   },
   "2026-09-26": {
    "move_pct": 0.36,
    "regime": "flat"
   },
   "2026-09-27": {
    "move_pct": 0.2,
    "regime": "flat"
   }
  },
  "side_mix": [
   {
    "model": "claude-fable-5",
    "n": 1330,
    "long": 0.4376,
    "short": 0.1158,
    "sideways": 0.4466,
    "ls_ratio": 3.78
   },
   {
    "model": "claude-opus-5",
    "n": 1330,
    "long": 0.3647,
    "short": 0.1015,
    "sideways": 0.5338,
    "ls_ratio": 3.59
   },
   {
    "model": "deepseek-v4-pro",
    "n": 1310,
    "long": 0.4153,
    "short": 0.1779,
    "sideways": 0.4069,
    "ls_ratio": 2.33
   },
   {
    "model": "gemini-3.1-pro",
    "n": 1330,
    "long": 0.4113,
    "short": 0.1857,
    "sideways": 0.403,
    "ls_ratio": 2.21
   },
   {
    "model": "gpt-5.6-sol",
    "n": 1330,
    "long": 0.3714,
    "short": 0.188,
    "sideways": 0.4406,
    "ls_ratio": 1.98
   },
   {
    "model": "grok-4.6",
    "n": 1323,
    "long": 0.263,
    "short": 0.1429,
    "sideways": 0.5941,
    "ls_ratio": 1.84
   },
   {
    "model": "qwen-3.8-max",
    "n": 1320,
    "long": 0.4098,
    "short": 0.1583,
    "sideways": 0.4318,
    "ls_ratio": 2.59
   }
  ],
  "long_short_ratio_field": {
   "value": 2.62,
   "method": "mean of the per-model long-share / short-share ratios"
  },
  "consensus_skew": {
   "cells_ge5_same_side": 606,
   "n_mature": 3902,
   "hit_rate": 0.4646
  },
  "fail_of_week": {
   "model": "gemini-3.1-pro",
   "symbol": "BTC",
   "side": "long",
   "confidence": 80,
   "slot": "2026-09-23T00:01:00+00:00",
   "fh": "1d",
   "tf": "1d",
   "exit_reason": "sl",
   "pnl_net_usd": -3.1
  },
  "counters": {
   "forecasts_total": 10290,
   "ok_in_gate": 9273,
   "out_of_gate": 973,
   "out_of_gate_by_fh": {
    "1w": 488,
    "1m": 485
   },
   "invalid": 44,
   "mature": 9273,
   "mature_directional": 4958,
   "mature_sideways": 4315,
   "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": {
    "2026-09-21": {
     "move_pct": 6.39,
     "regime": "trend"
    },
    "2026-09-22": {
     "move_pct": -0.4,
     "regime": "flat"
    },
    "2026-09-23": {
     "move_pct": -2.04,
     "regime": "trend"
    },
    "2026-09-24": {
     "move_pct": 0.02,
     "regime": "flat"
    },
    "2026-09-25": {
     "move_pct": -0.46,
     "regime": "flat"
    },
    "2026-09-26": {
     "move_pct": 0.36,
     "regime": "flat"
    },
    "2026-09-27": {
     "move_pct": 0.2,
     "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": "2026-09-21T00:01:00+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 -- four 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 #7) 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": [
      "2026-09-21",
      "2026-09-22",
      "2026-09-23",
      "2026-09-24",
      "2026-09-25",
      "2026-09-26",
      "2026-09-27"
     ],
     "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": [
      "2026-09-21",
      "2026-09-22",
      "2026-09-23",
      "2026-09-24",
      "2026-09-25",
      "2026-09-26",
      "2026-09-27"
     ],
     "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": [
      "2026-09-21",
      "2026-09-22",
      "2026-09-23",
      "2026-09-24",
      "2026-09-25",
      "2026-09-26",
      "2026-09-27"
     ],
     "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": [
      "2026-09-21",
      "2026-09-22",
      "2026-09-23",
      "2026-09-24",
      "2026-09-25",
      "2026-09-26",
      "2026-09-27"
     ],
     "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": [
      "2026-09-21",
      "2026-09-22",
      "2026-09-23",
      "2026-09-24",
      "2026-09-25",
      "2026-09-26",
      "2026-09-27"
     ],
     "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": {
   "qwen-3.8-max": 0.9891,
   "claude-fable-5": 0.9993,
   "claude-opus-5": 1.0,
   "grok-4.6": 0.9952,
   "deepseek-v4-pro": 0.9864,
   "gemini-3.1-pro": 1.0,
   "gpt-5.6-sol": 1.0
  },
  "calibration_bridge": {
   "best_calibrated_model": "qwen-3.8-max",
   "brier": 0.2689,
   "hit_rate_leader": "gemini-3.1-pro"
  }
 },
 "key_findings": [
  "**OBSERVATION -- No weekly title again.** gemini-3.1-pro leads descriptively at 51.4% [47.9%, 54.9%]; the gap to #2 (claude-opus-5, 47.6%) is 3.8pp. 0 of 7 models improved week over week; issue #7's leader gemini-3.1-pro is now 1 of 7.",
  "**No ticker cleared 51%.** XRP is the week's most readable coin (50.0%, n=1,165) and BNB the hardest at 38.2%; 0 of the 5 assets cleared 51%. Pair of the week: gemini-3.1-pro x SOL 54.3% (n=186); worst: claude-fable-5 x BNB 36.0% (n=114).",
  "**2 of 2 qualifying turns found a caller.** BTC short -2.90%, 1 caller(s), first gemini-3.1-pro; ETH short -3.73%, 6 caller(s), first gpt-5.6-sol.",
  "**Self-agreement split across the four cells.** Cross-TF/FH lifts ran -2.5pp to +3.8pp (issue #7: +1.4pp to +6.5pp). On trend days, 4h TF-agreement hit 64.0% (n=264) against 27.7% (n=314) on flat days."
 ],
 "market_check": {
  "label": "MARKET CHECK · OBSERVATION (two adjacent calendar weeks)",
  "text": "OBSERVATION — mean |1d move| 2.69% -> 1.83% (-32% rel), BTC realized vol 35.1% -> 35.8% (ann., hourly); field directional accuracy 50.7% -> 46.8% (-3.9pp), 0 of 7 models improved, sim win-rate up for 0 of 7, field sim PnL -$337.72 -> -$315.68 (adjacent calendar weeks Sep 14-20 vs Sep 21-27; descriptive, one pair of weeks, not a claim).",
  "robustness": "Robustness: raw price-sign accuracy 54.2% -> 49.2% — the same direction as the trade-based hit rule.",
  "weeks": {
   "A": {
    "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"
   },
   "B": {
    "days": "2026-09-21..2026-09-27",
    "slots": [
     "2026-09-21T00:00:00+00:00",
     "2026-09-28T00:00:00+00:00"
    ],
    "cutoff": "2026-09-28T16:00:00+00:00"
   }
  },
  "price_source": "crypto_spot/1h@:00",
  "accuracy_field": {
   "A": {
    "n": 5598,
    "hit_rate": 0.5075,
    "wilson_95": [
     0.4944,
     0.5206
    ]
   },
   "B": {
    "n": 4958,
    "hit_rate": 0.4683,
    "wilson_95": [
     0.4545,
     0.4822
    ]
   },
   "delta_pp": -3.9
  },
  "accuracy_per_model": [
   {
    "model": "gemini-3.1-pro",
    "model_label": "gemini-3.1-pro",
    "A": {
     "n": 883,
     "hit_rate": 0.5289,
     "wilson_95": [
      0.4959,
      0.5616
     ]
    },
    "B": {
     "n": 794,
     "hit_rate": 0.5139,
     "wilson_95": [
      0.4791,
      0.5485
     ]
    },
    "delta_pp": -1.5
   },
   {
    "model": "grok-4.6",
    "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "A": {
     "n": 629,
     "hit_rate": 0.496,
     "wilson_95": [
      0.4571,
      0.535
     ]
    },
    "B": {
     "n": 537,
     "hit_rate": 0.4693,
     "wilson_95": [
      0.4274,
      0.5116
     ]
    },
    "delta_pp": -2.7
   },
   {
    "model": "gpt-5.6-sol",
    "model_label": "gpt-5.6-sol",
    "A": {
     "n": 821,
     "hit_rate": 0.497,
     "wilson_95": [
      0.4628,
      0.5311
     ]
    },
    "B": {
     "n": 744,
     "hit_rate": 0.4677,
     "wilson_95": [
      0.4321,
      0.5037
     ]
    },
    "delta_pp": -2.9
   },
   {
    "model": "claude-opus-5",
    "model_label": "claude-opus-5",
    "A": {
     "n": 711,
     "hit_rate": 0.5162,
     "wilson_95": [
      0.4795,
      0.5527
     ]
    },
    "B": {
     "n": 620,
     "hit_rate": 0.4758,
     "wilson_95": [
      0.4368,
      0.5151
     ]
    },
    "delta_pp": -4.0
   },
   {
    "model": "deepseek-v4-pro",
    "model_label": "deepseek-v4-pro",
    "A": {
     "n": 936,
     "hit_rate": 0.484,
     "wilson_95": [
      0.4521,
      0.516
     ]
    },
    "B": {
     "n": 777,
     "hit_rate": 0.4389,
     "wilson_95": [
      0.4044,
      0.474
     ]
    },
    "delta_pp": -4.5
   },
   {
    "model": "claude-fable-5",
    "model_label": "claude-fable-5",
    "A": {
     "n": 799,
     "hit_rate": 0.5232,
     "wilson_95": [
      0.4885,
      0.5576
     ]
    },
    "B": {
     "n": 736,
     "hit_rate": 0.4674,
     "wilson_95": [
      0.4316,
      0.5035
     ]
    },
    "delta_pp": -5.6
   },
   {
    "model": "qwen-3.8-max",
    "model_label": "qwen-3.8-max",
    "A": {
     "n": 819,
     "hit_rate": 0.5079,
     "wilson_95": [
      0.4737,
      0.5421
     ]
    },
    "B": {
     "n": 750,
     "hit_rate": 0.4453,
     "wilson_95": [
      0.4101,
      0.4811
     ]
    },
    "delta_pp": -6.3
   }
  ],
  "raw_sign_field": {
   "A": {
    "n_scored": 5583,
    "hit_rate": 0.5416
   },
   "B": {
    "n_scored": 4954,
    "hit_rate": 0.4923
   }
  },
  "abs_move_1d": {
   "A": 2.693,
   "B": 1.835,
   "delta_pct_rel": -31.9
  },
  "btc_realized_vol_ann_pct_hourly": {
   "A": 35.07,
   "B": 35.8
  },
  "trading_field": {
   "A": {
    "n_trades": 5598,
    "wr": 0.4337,
    "pnl_net_usd": -337.72,
    "pnl_gross_usd": 222.08
   },
   "B": {
    "n_trades": 4958,
    "wr": 0.3998,
    "pnl_net_usd": -315.68,
    "pnl_gross_usd": 180.12
   }
  },
  "verdict": {
   "models_with_hit_improved": "0 of 7",
   "models_with_rawsign_improved": "0 of 7",
   "models_with_wr_improved": "0 of 7",
   "field_hit_delta_pp": -3.9,
   "field_pnl_net": {
    "A": -337.72,
    "B": -315.68
   },
   "abs_move_delta": {
    "1h": {
     "A": 0.369,
     "B": 0.358,
     "delta_pct_rel": -3.0
    },
    "4h": {
     "A": 0.835,
     "B": 0.766,
     "delta_pct_rel": -8.3
    },
    "1d": {
     "A": 2.693,
     "B": 1.835,
     "delta_pct_rel": -31.9
    }
   },
   "btc_vol_ann_pct": {
    "A": 35.07,
    "B": 35.8
   },
   "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, four windows back, inside issue #4's window). Week A (2026-09-14..2026-09-20) is pure grok-4.6 (it reproduces the published issue #7); 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 #7 as well -- the 4.5 -> 4.6 flip sits four 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 four 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 third 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. Eight issues, and the best ticker changed hands in every one except issue #6 (SOL -> BNB -> BTC -> SOL -> BNB -> BNB -> ETH -> XRP): XRP leads this week at 50.0%, after one issue of ETH.",
  "Self-agreement lifts run -2.5pp to +3.8pp with small disagree cells; carry the regime split into the monthly test, do not act on it."
 ],
 "limitations": [
  "Single week (Sep 21-27, 2026 UTC). Descriptive for this window only; no claim about next week or any model's underlying skill. This window ran 2 trend days of 7 (issue #7: 3 of 7); every comparison with issue #7 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. 15 events across eight issues cannot characterise detector performance either way.",
  "Cross-TF and cross-FH disagreement cells are small (n=21, 8, 25, 9); 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 routing of V4 Pro to V4.1 Flash from Sep 14, 04:00 UTC; an API check on Sep 29 still returns deepseek-v4-pro as the served model for our requests, with the same system fingerprint as on Sep 22, so this issue keeps the label; it will change the day the served id changes. On Sep 21, from 04:01 UTC, the provider returned HTTP 402 (insufficient balance) to 20 deepseek-v4-pro calls inside this window; they are counted as invalid (invalid 44 this window; the line's ok-rate 98.6%). Series density and lineage notes (wave 8): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Sep 21-27) daily coverage is FULL, as it was in issues #4, #5, #6 and #7: 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, four 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-#7 window), so every grok number in this issue and in issue #7 is a pure grok-4.6 line -- the grok week-over-week row is the third 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 #7 after the audit that found the raw table mixes two exchanges; issue #7 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 28 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4, #5, #6 and #7, which printed 52,971 at the Sep 21 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 57,929 is an additive step under one definition; counter deltas against issue #6 and earlier remain definitional."
 ],
 "notes": {
  "series_density_and_lineage_wave8": "Series density and lineage notes (wave 8): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Sep 21-27) daily coverage is FULL, as it was in issues #4, #5, #6 and #7: 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, four 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-#7 window), so every grok number in this issue and in issue #7 is a pure grok-4.6 line -- the grok week-over-week row is the third 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": "2026-09-21T00:01:00+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 -- four 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 #7) 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": [
     "2026-09-21",
     "2026-09-22",
     "2026-09-23",
     "2026-09-24",
     "2026-09-25",
     "2026-09-26",
     "2026-09-27"
    ],
    "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": [
     "2026-09-21",
     "2026-09-22",
     "2026-09-23",
     "2026-09-24",
     "2026-09-25",
     "2026-09-26",
     "2026-09-27"
    ],
    "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": [
     "2026-09-21",
     "2026-09-22",
     "2026-09-23",
     "2026-09-24",
     "2026-09-25",
     "2026-09-26",
     "2026-09-27"
    ],
    "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": [
     "2026-09-21",
     "2026-09-22",
     "2026-09-23",
     "2026-09-24",
     "2026-09-25",
     "2026-09-26",
     "2026-09-27"
    ],
    "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": [
     "2026-09-21",
     "2026-09-22",
     "2026-09-23",
     "2026-09-24",
     "2026-09-25",
     "2026-09-26",
     "2026-09-27"
    ],
    "missing_days": []
   }
  },
  "week_over_week_note": "Week-over-week columns compare back-to-back windows: Sep 14-20 (issue #7, and week A of this issue's alive slice) vs Sep 21-27 (this issue). \"Was\" values are the numbers published in issue #7; 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 28 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4, #5, #6 and #7, which printed 52,971 at the Sep 21 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 57,929 is an additive step under one definition; counter deltas against issue #6 and earlier remain definitional.",
  "market_state_audit": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #7 after the audit that found the raw table mixes two exchanges; issue #7 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 #8.** Weekly Model Watch is a living series. Issue #7's open questions -- \"does gemini-3.1-pro hold the top rank for a fourth issue, and does the title rule ever fire?\" and \"Does a qualifying reversal ever fire outside ETH, and does a second turn in the same pair ever draw a caller?\" -- read this issue as: gemini-3.1-pro at the top (51.4%), the line that also led issue #7 (gemini-3.1-pro), so the rank held for a fourth issue; no weekly title was awarded, for the eighth issue running, and 2 of 2 qualifying turns found a caller, on turns in BTC / ETH; the BTC turn is the first qualifying turn outside ETH since issue #4, and with one turn per pair the second-turn question stays open. The grok row is pure grok-4.6 this issue and was pure grok-4.6 in issue #7; the 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits four 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 fifth issue, and does the title rule ever fire?",
  "Does a qualifying reversal fire outside ETH for a second issue running, 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 #8",
   "url": "https://marketmania.ai/research/reports/consensus-watch-2026-09-21.pdf"
  },
  {
   "title": "Weekly Calibration #8",
   "url": "https://marketmania.ai/research/reports/weekly-calibration-2026-09-21.pdf"
  },
  {
   "title": "Weekly Model Watch #7",
   "url": "https://marketmania.ai/research/reports/model-watch-2026-09-14.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": 4958
  },
  "platform": {
   "as_of_cutoff": "2026-09-28",
   "resolved_forecasts": 57929,
   "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": 37
  },
  "source": "weekly_metrics_2026-09-21.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 28 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4, #5, #6 and #7, which printed 52,971 at the Sep 21 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 57,929 is an additive step under one definition; counter deltas against issue #6 and earlier remain definitional."
 },
 "citation": {
  "bibtex_key": "mm_modelwatch_2026w39",
  "title": "Weekly Model Watch #8: weekly leaderboard, ticker/pair reads and cross-confirmation, Sep 21-27 2026",
  "author": "MarketMania Research",
  "year": 2026,
  "month": "September",
  "day": 30,
  "url": "https://marketmania.ai/research/reports/model-watch-2026-09-21.pdf",
  "note": "Methodology v1.1; window Sep 21-27, 2026 UTC; source weekly_metrics_2026-09-21.json"
 },
 "site_entry": {
  "slug": "model-watch-2026-09-21",
  "series": "weekly",
  "seriesLabel": "Weekly · Model Watch",
  "date": "September 30, 2026",
  "title": "Weekly Model Watch #8",
  "summary": "4,958 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 51.4% (-1.5pp w/w) against claude-opus-5 at 47.6%, a gap of 3.8pp. XRP was the most readable ticker at 50.0% and BNB the hardest at 38.2%, with 0 of 5 assets above 51 percent. Reversals: 2 of 2 qualifying turns found a caller. Field accuracy moved 50.7% -> 46.8% (-3.9pp) with 0 of 7 models improving.",
  "stats": [
   "4,958 directional calls",
   "No weekly title again",
   "field -3.9pp w/w"
  ],
  "manifest_key_finding": "Field accuracy moved 50.7% -> 46.8% (-3.9pp) with 0 of 7 models improving; gemini-3.1-pro led the week at 51.4% (-1.5pp w/w) and no weekly title was awarded, for the eighth issue running.",
  "files": {
   "pdf": "/research/reports/model-watch-2026-09-21.pdf",
   "md": "/research/reports/model-watch-2026-09-21.md",
   "json": "/research/reports/model-watch-2026-09-21.json"
  }
 },
 "social": {
  "telegram": {
   "photo_caption_html": "📊 <b>Weekly Model Watch #8</b> — weekly series (window Sep 21–27)\n\nWho led, what was readable, who saw the turn, does self-agreement help? 4,958 directional calls, 7 model lines:\n• <b>No weekly title again</b> — gemini-3.1-pro leads at <b>51.4%</b> (-1.5pp w/w) against claude-opus-5 at 47.6%, a 3.8pp gap\n• Tickers: <b>XRP</b> most readable (50.0%); BNB hardest at 38.2% — 0 of 5 assets above 51%\n• Reversals: 2 of 2 qualifying turns found a caller\n• Market check: field accuracy 50.7% -> 46.8% (-3.9pp), 0 of 7 improved; |1d| move 2.69% -> 1.83%\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 four windows back\n\n📄 <a href=\"https://marketmania.ai/research/reports/model-watch-2026-09-21.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 #8 — full report (PDF).\nWeb copy: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-09-21.pdf\">model-watch-2026-09-21.pdf</a> · machine-readable: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-09-21.json\">JSON</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #models #research",
   "caption_len": 892,
   "document_caption_len": 390,
   "limit": 1024
  },
  "x": {
   "main": "Weekly Model Watch #8 is out.\n\n4,958 directional LLM calls: field accuracy 50.7% -> 46.8% w/w, 0 of 7 models improved. gemini-3.1-pro led at 51.4%; XRP was the most readable ticker.\n\nFull report: https://marketmania.ai/research/reports/model-watch-2026-09-21.pdf\n\n#AI #Crypto #Research #LLM",
   "main_tco_len": 247,
   "reply": "Hit-rate, w/w:\ngemini-3.1: 52.9 -> 51.4\nopus-5: 51.6 -> 47.6\ngrok-4.6: 49.6 -> 46.9\ngpt-5.6: 49.7 -> 46.8\nfable-5: 52.3 -> 46.7\nqwen-3.8: 50.8 -> 44.5\ndeepseek: 48.4 -> 43.9\nField: 50.7 -> 46.8. Wilson CIs; methodology v1.1.\nData: https://marketmania.ai/research/reports/model-watch-2026-09-21.json",
   "reply_tco_len": 254,
   "limit_tco": 280,
   "replies_max": 1
  }
 }
}