{
 "slug": "model-watch-2026-08-31",
 "series": "weekly",
 "series_label": "WEEKLY · MODEL WATCH",
 "issue": 5,
 "title": "Weekly Model Watch #5",
 "language": "en",
 "window_start": "2026-08-31T00:00:00+00:00",
 "window_end": "2026-09-07T00:00:00+00:00",
 "window_text": "Aug 31-Sep 6, 2026 UTC",
 "cutoff": "2026-09-07T16:00:00+00:00",
 "generated_at": "2026-09-08T09:12:03+00:00",
 "methodology": "v1.1 (2026-08-10)",
 "methodology_hash": "e66c7e8c864a2233",
 "source": "weekly_metrics_2026-08-31.json",
 "hit_rule": "direction: exit tp1/tp2 -> hit, sl -> miss, expiry -> sign of gross pnl",
 "bibtex_key": "mm_modelwatch_2026w36",
 "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": 4520,
  "hit_rate": 0.4283
 },
 "market_state": {
  "panel": "market_state_6p2",
  "week": {
   "start": "2026-08-31",
   "end": "2026-09-07"
  },
  "prev_week": {
   "start": "2026-08-24",
   "end": "2026-08-31"
  },
  "current": {
   "btc_net_pct": 3.42,
   "realized_vol_ann_pct": 43.35,
   "volume_top5_usd_bn": 16.2391,
   "volume_top5_wow_pct": -17.7,
   "avg_pairwise_corr": 0.78
  },
  "previous_as_published_issue4": {
   "btc_net_pct": -0.07,
   "realized_vol_ann_pct": 30.85,
   "volume_top5_usd_bn": 19.7224,
   "avg_pairwise_corr": 0.85
  },
  "snapshot_line": "BTC net +3.42% (prior -0.07%) - ann vol 43.4% (was 30.9%) - TOP5 volume $16.24B, -18% w/w - pairwise corr 0.78 (was 0.85)",
  "exchange": "binance",
  "audit_filter_applied": true,
  "source": "single-exchange (binance) daily candle recompute, market_state_2026-08-31.json; prior-week values as published in issue #4",
  "audit_note": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #4 after the audit that found the raw table mixes two exchanges; issue #4 values are as published.",
  "prev_week_control": {
   "published_issue3": {
    "btc_net_pct": -0.07,
    "realized_vol_ann_pct": 30.85,
    "volume_top5_usd_bn": 19.7224,
    "avg_pairwise_corr": 0.85
   },
   "published_issue4": {
    "btc_net_pct": -0.07,
    "realized_vol_ann_pct": 30.85,
    "volume_top5_usd_bn": 19.7224,
    "avg_pairwise_corr": 0.85
   },
   "got": {
    "btc_net_pct": -0.07,
    "realized_vol_ann_pct": 30.85,
    "volume_top5_usd_bn": 19.7224,
    "avg_pairwise_corr": 0.85
   },
   "match": {
    "btc_net_pct": true,
    "realized_vol_ann_pct": true,
    "volume_top5_usd_bn": true,
    "avg_pairwise_corr": true
   },
   "tolerance": {
    "btc_net_pct": 0.05,
    "realized_vol_ann_pct": 0.5,
    "volume_top5_usd_bn": 0.05,
    "avg_pairwise_corr": 0.02
   },
   "all_match": true,
   "note": "previous week of this run == published issue-#4 week; the four Snapshot pins must reproduce (key published_issue3 kept for the builder schema, published_issue4 is the add-only alias)"
  },
  "regime_days": {
   "2026-08-31": {
    "move_pct": 1.19,
    "regime": "flat"
   },
   "2026-09-01": {
    "move_pct": -1.5,
    "regime": "trend"
   },
   "2026-09-02": {
    "move_pct": -0.06,
    "regime": "flat"
   },
   "2026-09-03": {
    "move_pct": 4.98,
    "regime": "trend"
   },
   "2026-09-04": {
    "move_pct": -1.97,
    "regime": "trend"
   },
   "2026-09-05": {
    "move_pct": 0.24,
    "regime": "flat"
   },
   "2026-09-06": {
    "move_pct": 0.58,
    "regime": "flat"
   }
  },
  "alive_slice_context": {
   "abs_move_1d_pct": {
    "A": 2.352,
    "B": 2.102
   },
   "btc_realized_vol_ann_pct_hourly": {
    "A": 37.9,
    "B": 35.22
   }
  }
 },
 "key_finding": {
  "label": "KEY FINDING · OBSERVATION (one weekly window)",
  "text": "Field accuracy went 42.9% -> 42.8% (-0.1pp) with 1 of 7 models improving, and no weekly title was awarded, for the fifth issue running.",
  "evidence_level": "observation"
 },
 "tables": {
  "leaderboard": {
   "rows": [
    {
     "model": "gemini-3.1-pro",
     "n": 761,
     "hit_rate": 0.4652,
     "wilson_95": [
      0.43,
      0.5007
     ],
     "coverage": 0.5726,
     "model_label": "gemini-3.1-pro",
     "insufficient": false
    },
    {
     "model": "deepseek-v4-pro",
     "n": 724,
     "hit_rate": 0.442,
     "wilson_95": [
      0.4062,
      0.4784
     ],
     "coverage": 0.5464,
     "model_label": "deepseek-v4-pro",
     "insufficient": false
    },
    {
     "model": "claude-fable-5",
     "n": 628,
     "hit_rate": 0.4283,
     "wilson_95": [
      0.3902,
      0.4674
     ],
     "coverage": 0.4722,
     "model_label": "claude-fable-5",
     "insufficient": false
    },
    {
     "model": "qwen-3.8-max",
     "n": 697,
     "hit_rate": 0.4247,
     "wilson_95": [
      0.3885,
      0.4617
     ],
     "coverage": 0.5353,
     "model_label": "qwen-3.8-max",
     "insufficient": false
    },
    {
     "model": "claude-opus-5",
     "n": 545,
     "hit_rate": 0.4239,
     "wilson_95": [
      0.383,
      0.4657
     ],
     "coverage": 0.411,
     "model_label": "claude-opus-5",
     "insufficient": false
    },
    {
     "model": "gpt-5.6-sol",
     "n": 689,
     "hit_rate": 0.4165,
     "wilson_95": [
      0.3803,
      0.4537
     ],
     "coverage": 0.518,
     "model_label": "gpt-5.6-sol",
     "insufficient": false
    },
    {
     "model": "grok-4.6",
     "n": 476,
     "hit_rate": 0.3761,
     "wilson_95": [
      0.3337,
      0.4204
     ],
     "coverage": 0.3601,
     "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
     "insufficient": false
    }
   ],
   "title_awarded": false,
   "title_rule": ">=5pp gap AND non-overlapping 95% CI, N>=10",
   "top2_gap_pp": 2.3,
   "movement_vs_prior_week": [
    {
     "model": "gemini-3.1-pro",
     "hit_rate_issue4_published": 0.4424,
     "hit_rate_issue5": 0.4652,
     "delta_pp": 2.3,
     "rank_issue5": 1
    },
    {
     "model": "deepseek-v4-pro",
     "hit_rate_issue4_published": 0.4458,
     "hit_rate_issue5": 0.442,
     "delta_pp": -0.4,
     "rank_issue5": 2
    },
    {
     "model": "claude-fable-5",
     "hit_rate_issue4_published": 0.4301,
     "hit_rate_issue5": 0.4283,
     "delta_pp": -0.2,
     "rank_issue5": 3
    },
    {
     "model": "qwen-3.8-max",
     "hit_rate_issue4_published": 0.4256,
     "hit_rate_issue5": 0.4247,
     "delta_pp": -0.1,
     "rank_issue5": 4
    },
    {
     "model": "claude-opus-5",
     "hit_rate_issue4_published": 0.4242,
     "hit_rate_issue5": 0.4239,
     "delta_pp": -0.0,
     "rank_issue5": 5
    },
    {
     "model": "gpt-5.6-sol",
     "hit_rate_issue4_published": 0.4266,
     "hit_rate_issue5": 0.4165,
     "delta_pp": -1.0,
     "rank_issue5": 6
    },
    {
     "model": "grok-4.6",
     "hit_rate_issue4_published": 0.3987,
     "hit_rate_issue5": 0.3761,
     "delta_pp": -2.3,
     "rank_issue5": 7
    }
   ],
   "note": "No weekly title is awarded this issue -- for the fifth time. Title rule: **>=5pp gap AND non-overlapping 95% CI, N>=10** per side. gemini-3.1-pro's lead over deepseek-v4-pro is 2.3pp and their CIs overlap, so neither condition is met."
  },
  "ticker": {
   "rows": [
    {
     "symbol": "BNB",
     "n": 916,
     "hit_rate": 0.4727,
     "wilson_95": [
      0.4406,
      0.5051
     ],
     "issue4_hit_rate": 0.4123
    },
    {
     "symbol": "XRP",
     "n": 909,
     "hit_rate": 0.4422,
     "wilson_95": [
      0.4103,
      0.4747
     ],
     "issue4_hit_rate": 0.4027
    },
    {
     "symbol": "ETH",
     "n": 982,
     "hit_rate": 0.4308,
     "wilson_95": [
      0.4001,
      0.4619
     ],
     "issue4_hit_rate": 0.4329
    },
    {
     "symbol": "BTC",
     "n": 780,
     "hit_rate": 0.4064,
     "wilson_95": [
      0.3725,
      0.4413
     ],
     "issue4_hit_rate": 0.4372
    },
    {
     "symbol": "SOL",
     "n": 933,
     "hit_rate": 0.3869,
     "wilson_95": [
      0.3562,
      0.4186
     ],
     "issue4_hit_rate": 0.4576
    }
   ],
   "best": "BNB",
   "hardest": "SOL",
   "all_above_51pct": false
  },
  "pairs_top": [
   {
    "model": "deepseek-v4-pro",
    "symbol": "BNB",
    "n": 144,
    "hit_rate": 0.5347
   },
   {
    "model": "claude-opus-5",
    "symbol": "BNB",
    "n": 112,
    "hit_rate": 0.4911
   },
   {
    "model": "claude-fable-5",
    "symbol": "BNB",
    "n": 127,
    "hit_rate": 0.4803
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "BNB",
    "n": 155,
    "hit_rate": 0.4774
   },
   {
    "model": "deepseek-v4-pro",
    "symbol": "ETH",
    "n": 164,
    "hit_rate": 0.4695
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "XRP",
    "n": 147,
    "hit_rate": 0.4694
   }
  ],
  "pairs_worst": [
   {
    "model": "claude-opus-5",
    "symbol": "SOL",
    "n": 112,
    "hit_rate": 0.3661
   },
   {
    "model": "deepseek-v4-pro",
    "symbol": "SOL",
    "n": 148,
    "hit_rate": 0.3649
   },
   {
    "model": "grok-4.6",
    "symbol": "BTC",
    "n": 70,
    "hit_rate": 0.3571
   },
   {
    "model": "grok-4.6",
    "symbol": "ETH",
    "n": 99,
    "hit_rate": 0.3535
   },
   {
    "model": "grok-4.6",
    "symbol": "SOL",
    "n": 105,
    "hit_rate": 0.2952
   }
  ],
  "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": "2026-09-03T08:00:00+00:00",
     "new_side": "long",
     "move_8h_pct": 4.61,
     "n_callers": 0,
     "first_caller": null,
     "caller_models": []
    },
    {
     "symbol": "ETH",
     "t": "2026-09-04T08:00:00+00:00",
     "new_side": "short",
     "move_8h_pct": -2.13,
     "n_callers": 0,
     "first_caller": null,
     "caller_models": []
    }
   ],
   "scope_v1": "BTC/ETH only",
   "status": "ok"
  },
  "cross_tf": {
   "fh_4h_tf_4h_vs_1h": {
    "agree": {
     "n": 461,
     "hit_rate": 0.3536,
     "wilson_95": [
      0.3113,
      0.3983
     ],
     "by_regime": {
      "trend": {
       "n": 214,
       "hit_rate": 0.3364
      },
      "flat": {
       "n": 247,
       "hit_rate": 0.3684
      }
     }
    },
    "disagree": {
     "n": 10,
     "hit_rate": 0.9,
     "wilson_95": [
      0.5958,
      0.9821
     ],
     "by_regime": {
      "trend": {
       "n": 5,
       "hit_rate": 1.0
      },
      "flat": {
       "n": 5,
       "hit_rate": 0.8
      }
     }
    },
    "mixed_pairs": 177,
    "base": {
     "n": 651,
     "hit_rate": 0.3994
    },
    "lift_pp": -4.6
   },
   "fh_1d_tf_1d_vs_4h": {
    "agree": {
     "n": 67,
     "hit_rate": 0.2836,
     "wilson_95": [
      0.1897,
      0.4009
     ],
     "by_regime": {
      "trend": {
       "n": 38,
       "hit_rate": 0.0526
      },
      "flat": {
       "n": 29,
       "hit_rate": 0.5862
      }
     }
    },
    "disagree": {
     "n": 12,
     "hit_rate": 0.8333,
     "wilson_95": [
      0.552,
      0.953
     ],
     "by_regime": {
      "trend": {
       "n": 0,
       "hit_rate": null
      },
      "flat": {
       "n": 12,
       "hit_rate": 0.8333
      }
     }
    },
    "mixed_pairs": 49,
    "base": {
     "n": 128,
     "hit_rate": 0.4375
    },
    "lift_pp": -15.4
   }
  },
  "cross_fh": {
   "fh_4h_confirmed_by_1h": {
    "agree": {
     "n": 401,
     "hit_rate": 0.3741,
     "wilson_95": [
      0.3281,
      0.4224
     ],
     "by_regime": {
      "trend": {
       "n": 188,
       "hit_rate": 0.3723
      },
      "flat": {
       "n": 213,
       "hit_rate": 0.3756
      }
     }
    },
    "disagree": {
     "n": 20,
     "hit_rate": 0.75,
     "wilson_95": [
      0.5313,
      0.8881
     ],
     "by_regime": {
      "trend": {
       "n": 10,
       "hit_rate": 0.7
      },
      "flat": {
       "n": 10,
       "hit_rate": 0.8
      }
     }
    },
    "mixed_pairs": 227,
    "base": {
     "n": 651,
     "hit_rate": 0.3994
    },
    "lift_pp": -2.5
   },
   "fh_1d_confirmed_by_4h": {
    "agree": {
     "n": 57,
     "hit_rate": 0.2632,
     "wilson_95": [
      0.1665,
      0.3898
     ],
     "by_regime": {
      "trend": {
       "n": 35,
       "hit_rate": 0.0571
      },
      "flat": {
       "n": 22,
       "hit_rate": 0.5909
      }
     }
    },
    "disagree": {
     "n": 12,
     "hit_rate": 0.8333,
     "wilson_95": [
      0.552,
      0.953
     ],
     "by_regime": {
      "trend": {
       "n": 0,
       "hit_rate": null
      },
      "flat": {
       "n": 12,
       "hit_rate": 0.8333
      }
     }
    },
    "mixed_pairs": 59,
    "base": {
     "n": 128,
     "hit_rate": 0.4375
    },
    "lift_pp": -17.4
   }
  },
  "regime_days": {
   "2026-08-31": {
    "move_pct": 1.19,
    "regime": "flat"
   },
   "2026-09-01": {
    "move_pct": -1.5,
    "regime": "trend"
   },
   "2026-09-02": {
    "move_pct": -0.06,
    "regime": "flat"
   },
   "2026-09-03": {
    "move_pct": 4.98,
    "regime": "trend"
   },
   "2026-09-04": {
    "move_pct": -1.97,
    "regime": "trend"
   },
   "2026-09-05": {
    "move_pct": 0.24,
    "regime": "flat"
   },
   "2026-09-06": {
    "move_pct": 0.58,
    "regime": "flat"
   }
  },
  "side_mix": [
   {
    "model": "claude-fable-5",
    "n": 1330,
    "long": 0.3263,
    "short": 0.1459,
    "sideways": 0.5278,
    "ls_ratio": 2.24
   },
   {
    "model": "claude-opus-5",
    "n": 1326,
    "long": 0.2949,
    "short": 0.1161,
    "sideways": 0.589,
    "ls_ratio": 2.54
   },
   {
    "model": "deepseek-v4-pro",
    "n": 1325,
    "long": 0.3125,
    "short": 0.234,
    "sideways": 0.4536,
    "ls_ratio": 1.34
   },
   {
    "model": "gemini-3.1-pro",
    "n": 1329,
    "long": 0.316,
    "short": 0.2566,
    "sideways": 0.4274,
    "ls_ratio": 1.23
   },
   {
    "model": "gpt-5.6-sol",
    "n": 1330,
    "long": 0.2639,
    "short": 0.2541,
    "sideways": 0.482,
    "ls_ratio": 1.04
   },
   {
    "model": "grok-4.6",
    "n": 1322,
    "long": 0.1536,
    "short": 0.2065,
    "sideways": 0.6399,
    "ls_ratio": 0.74
   },
   {
    "model": "qwen-3.8-max",
    "n": 1302,
    "long": 0.3402,
    "short": 0.1951,
    "sideways": 0.4647,
    "ls_ratio": 1.74
   }
  ],
  "long_short_ratio_field": {
   "value": 1.55,
   "method": "mean of the per-model long-share / short-share ratios"
  },
  "consensus_skew": {
   "cells_ge5_same_side": 521,
   "n_mature": 3248,
   "hit_rate": 0.4126
  },
  "fail_of_week": {
   "model": "gemini-3.1-pro",
   "symbol": "BTC",
   "side": "long",
   "confidence": 80,
   "slot": "2026-09-04T12:01:00+00:00",
   "fh": "4h",
   "tf": "4h",
   "exit_reason": "sl",
   "pnl_net_usd": -1.6
  },
  "counters": {
   "forecasts_total": 10290,
   "ok_in_gate": 9264,
   "out_of_gate": 980,
   "out_of_gate_by_fh": {
    "1w": 490,
    "1m": 490
   },
   "invalid": 46,
   "mature": 9264,
   "mature_directional": 4520,
   "mature_sideways": 4744,
   "pending_next_issue": 0,
   "pending_by_fh": {},
   "late_closes": 0,
   "uptime": [
    {
     "fh": "1h",
     "tf": "1h",
     "slots_seen": 168,
     "slots_expected": 168
    },
    {
     "fh": "4h",
     "tf": "4h",
     "slots_seen": 42,
     "slots_expected": 42
    },
    {
     "fh": "4h",
     "tf": "1h",
     "slots_seen": 42,
     "slots_expected": 42
    },
    {
     "fh": "1d",
     "tf": "1d",
     "slots_seen": 7,
     "slots_expected": 7
    },
    {
     "fh": "1d",
     "tf": "4h",
     "slots_seen": 7,
     "slots_expected": 7
    }
   ],
   "regime_days": {
    "2026-08-31": {
     "move_pct": 1.19,
     "regime": "flat"
    },
    "2026-09-01": {
     "move_pct": -1.5,
     "regime": "trend"
    },
    "2026-09-02": {
     "move_pct": -0.06,
     "regime": "flat"
    },
    "2026-09-03": {
     "move_pct": 4.98,
     "regime": "trend"
    },
    "2026-09-04": {
     "move_pct": -1.97,
     "regime": "trend"
    },
    "2026-09-05": {
     "move_pct": 0.24,
     "regime": "flat"
    },
    "2026-09-06": {
     "move_pct": 0.58,
     "regime": "flat"
    }
   },
   "grok_era": {
    "ids_in_window": [
     "grok-4.6"
    ],
    "n_grok45": 0,
    "n_grok46": 1470,
    "n_grok45_rows": 0,
    "n_grok46_rows": 1470,
    "first_grok46_slot": "2026-08-31T00:01:00+00:00",
    "last_grok45_slot": null,
    "first_grok45_slot": null,
    "flip_at": "2026-08-24T09:22:00+00:00",
    "flip_inside_window": false,
    "prev_window": {
     "n_grok45_rows": 110,
     "n_grok46_rows": 1360
    },
    "merged_model_id": "grok-4.6",
    "merged_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "lineage_ids": [
     "grok-4.5",
     "grok-4.6"
    ],
    "lineage_source": "benchmarks.config.lineage_ids('grok-4.6') + MODEL_SUCCESSION[grok-4.6]=('grok-4.5',)",
    "lineage_warnings": [],
    "rows_remapped_a_plus_b": 110,
    "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- inside the PREVIOUS window (week A); all per-model aggregates use the lineage splice grok-4.5 -> grok-4.6 (one row, labelled 'grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)'). Week B (this window) is pure grok-4.6; week A (prev window) mixes both eras and reproduces issue #4.",
    "cells_with_two_grok_rows": 0
   },
   "series_density": {
    "1d": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": true,
     "n_slots": 7,
     "n_rows": 490,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "1h": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": true,
     "n_slots": 168,
     "n_rows": 5880,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "1m": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": false,
     "n_slots": 7,
     "n_rows": 490,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "1w": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": false,
     "n_slots": 7,
     "n_rows": 490,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    },
    "4h": {
     "days_with_slots": 7,
     "expected_days": 7,
     "full_daily_coverage": true,
     "in_fh_gate": true,
     "n_slots": 42,
     "n_rows": 2940,
     "days": [
      "2026-08-31",
      "2026-09-01",
      "2026-09-02",
      "2026-09-03",
      "2026-09-04",
      "2026-09-05",
      "2026-09-06"
     ],
     "missing_days": []
    }
   },
   "model_roster": {
    "current": [
     "claude-fable-5",
     "claude-opus-5",
     "deepseek-v4-pro",
     "gemini-3.1-pro",
     "gpt-5.6-sol",
     "grok-4.6",
     "qwen-3.8-max"
    ],
    "current_count": 7,
    "current_raw_ids_in_window": [
     "claude-fable-5",
     "claude-opus-5",
     "deepseek-v4-pro",
     "gemini-3.1-pro",
     "gpt-5.6-sol",
     "grok-4.6",
     "qwen-3.8-max"
    ],
    "archived_legacy": [
     "grok-4.5",
     "qwen-3.7-max",
     "claude-opus-4.8"
    ],
    "archived_legacy_count": 3,
    "archived_legacy_detail": [
     {
      "model": "grok-4.5",
      "n_rows_since": 11128,
      "first_slot": "2026-07-15T18:01:00+00:00",
      "last_slot": "2026-08-24T09:01:00+00:00"
     },
     {
      "model": "qwen-3.7-max",
      "n_rows_since": 7479,
      "first_slot": "2026-07-15T18:01:00+00:00",
      "last_slot": "2026-08-05T09:01:00+00:00"
     },
     {
      "model": "claude-opus-4.8",
      "n_rows_since": 6315,
      "first_slot": "2026-07-15T18:01:00+00:00",
      "last_slot": "2026-07-30T08:01:00+00:00"
     }
    ],
    "archived_excluding_lineage_merged": [
     "qwen-3.7-max",
     "claude-opus-4.8"
    ],
    "tracked_total": 10,
    "canon_expected": {
     "current": 7,
     "archived_legacy": 4,
     "tracked_total": 11
    },
    "matches_canon_11_7_4": false,
    "roster_since": "2026-07-11T00:00:00+00:00",
    "note": "current = merged-lineage model ids emitting inside the window (grok-4.5 rows are spliced into grok-4.6); archived legacy = raw ids seen since 2026-07-11 that no longer emit in the window -- grok-4.5 is one of them by raw id even though its rows are spliced. Models retired before 2026-07-11 are not visible to this query."
   }
  },
  "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": 0.981,
   "deepseek-v4-pro": 0.9966,
   "claude-fable-5": 1.0,
   "claude-opus-5": 0.9973,
   "grok-4.6": 0.9946,
   "gemini-3.1-pro": 0.9993,
   "gpt-5.6-sol": 1.0
  },
  "calibration_bridge": {
   "best_calibrated_model": "qwen-3.8-max",
   "brier": 0.2755,
   "hit_rate_leader": "gemini-3.1-pro"
  }
 },
 "key_findings": [
  "**OBSERVATION -- No weekly title again.** gemini-3.1-pro leads descriptively at 46.5% [43.0%, 50.1%]; the gap to #2 (deepseek-v4-pro, 44.2%) is 2.3pp. 1 of 7 models improved week over week; issue #4's leader deepseek-v4-pro is now 2 of 7.",
  "**No ticker cleared 51%.** BNB is the week's most readable coin (47.3%, n=916) and SOL the hardest at 38.7%; 0 of the 5 assets cleared 51%. Pair of the week: deepseek-v4-pro x BNB 53.5% (n=144); worst: grok-4.6 x SOL 29.5% (n=105).",
  "**0 of 2 qualifying turns found a caller.** ETH long +4.61%, 0 caller(s), first NOBODY; ETH short -2.13%, 0 caller(s), first NOBODY.",
  "**Self-agreement read as a penalty in every cell.** Cross-TF/FH lifts ran -17.4pp to -2.5pp (issue #4: -7.6pp to -1.9pp). On trend days, 4h TF-agreement hit 33.6% (n=214) against 36.8% (n=247) on flat days."
 ],
 "market_check": {
  "label": "MARKET CHECK · OBSERVATION (two adjacent calendar weeks)",
  "text": "OBSERVATION — mean |1d move| 2.35% -> 2.10% (-11% rel), BTC realized vol 37.9% -> 35.2% (ann., hourly); field directional accuracy 42.9% -> 42.8% (-0.1pp), 1 of 7 models improved, sim win-rate up for 1 of 7, field sim PnL -$473.59 -> -$669.03 (adjacent calendar weeks Aug 24-30 vs Aug 31-Sep 6; descriptive, one pair of weeks, not a claim).",
  "robustness": "Robustness: raw price-sign accuracy 44.2% -> 51.6% — the opposite direction to the trade-based hit rule.",
  "weeks": {
   "A": {
    "days": "2026-08-24..2026-08-30",
    "slots": [
     "2026-08-24T00:00:00+00:00",
     "2026-08-31T00:00:00+00:00"
    ],
    "cutoff": "2026-08-31T16:00:00+00:00"
   },
   "B": {
    "days": "2026-08-31..2026-09-06",
    "slots": [
     "2026-08-31T00:00:00+00:00",
     "2026-09-07T00:00:00+00:00"
    ],
    "cutoff": "2026-09-07T16:00:00+00:00"
   }
  },
  "price_source": "crypto_spot/1h@:00",
  "accuracy_field": {
   "A": {
    "n": 4576,
    "hit_rate": 0.4294,
    "wilson_95": [
     0.4151,
     0.4438
    ]
   },
   "B": {
    "n": 4520,
    "hit_rate": 0.4283,
    "wilson_95": [
     0.414,
     0.4428
    ]
   },
   "delta_pp": -0.1
  },
  "accuracy_per_model": [
   {
    "model": "gemini-3.1-pro",
    "model_label": "gemini-3.1-pro",
    "A": {
     "n": 755,
     "hit_rate": 0.4424,
     "wilson_95": [
      0.4073,
      0.478
     ]
    },
    "B": {
     "n": 761,
     "hit_rate": 0.4652,
     "wilson_95": [
      0.43,
      0.5007
     ]
    },
    "delta_pp": 2.3
   },
   {
    "model": "claude-opus-5",
    "model_label": "claude-opus-5",
    "A": {
     "n": 561,
     "hit_rate": 0.4242,
     "wilson_95": [
      0.384,
      0.4655
     ]
    },
    "B": {
     "n": 545,
     "hit_rate": 0.4239,
     "wilson_95": [
      0.383,
      0.4657
     ]
    },
    "delta_pp": -0.0
   },
   {
    "model": "qwen-3.8-max",
    "model_label": "qwen-3.8-max",
    "A": {
     "n": 672,
     "hit_rate": 0.4256,
     "wilson_95": [
      0.3887,
      0.4633
     ]
    },
    "B": {
     "n": 697,
     "hit_rate": 0.4247,
     "wilson_95": [
      0.3885,
      0.4617
     ]
    },
    "delta_pp": -0.1
   },
   {
    "model": "claude-fable-5",
    "model_label": "claude-fable-5",
    "A": {
     "n": 651,
     "hit_rate": 0.4301,
     "wilson_95": [
      0.3926,
      0.4684
     ]
    },
    "B": {
     "n": 628,
     "hit_rate": 0.4283,
     "wilson_95": [
      0.3902,
      0.4674
     ]
    },
    "delta_pp": -0.2
   },
   {
    "model": "deepseek-v4-pro",
    "model_label": "deepseek-v4-pro",
    "A": {
     "n": 729,
     "hit_rate": 0.4458,
     "wilson_95": [
      0.4101,
      0.4821
     ]
    },
    "B": {
     "n": 724,
     "hit_rate": 0.442,
     "wilson_95": [
      0.4062,
      0.4784
     ]
    },
    "delta_pp": -0.4
   },
   {
    "model": "gpt-5.6-sol",
    "model_label": "gpt-5.6-sol",
    "A": {
     "n": 729,
     "hit_rate": 0.4266,
     "wilson_95": [
      0.3912,
      0.4628
     ]
    },
    "B": {
     "n": 689,
     "hit_rate": 0.4165,
     "wilson_95": [
      0.3803,
      0.4537
     ]
    },
    "delta_pp": -1.0
   },
   {
    "model": "grok-4.6",
    "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "A": {
     "n": 479,
     "hit_rate": 0.3987,
     "wilson_95": [
      0.3559,
      0.4432
     ]
    },
    "B": {
     "n": 476,
     "hit_rate": 0.3761,
     "wilson_95": [
      0.3337,
      0.4204
     ]
    },
    "delta_pp": -2.3
   }
  ],
  "raw_sign_field": {
   "A": {
    "n_scored": 4567,
    "hit_rate": 0.4419
   },
   "B": {
    "n_scored": 4507,
    "hit_rate": 0.5156
   }
  },
  "abs_move_1d": {
   "A": 2.352,
   "B": 2.102,
   "delta_pct_rel": -10.6
  },
  "btc_realized_vol_ann_pct_hourly": {
   "A": 37.9,
   "B": 35.22
  },
  "trading_field": {
   "A": {
    "n_trades": 4576,
    "wr": 0.3702,
    "pnl_net_usd": -473.59,
    "pnl_gross_usd": -15.99
   },
   "B": {
    "n_trades": 4520,
    "wr": 0.3564,
    "pnl_net_usd": -669.03,
    "pnl_gross_usd": -217.03
   }
  },
  "verdict": {
   "models_with_hit_improved": "1 of 7",
   "models_with_rawsign_improved": "7 of 7",
   "models_with_wr_improved": "1 of 7",
   "field_hit_delta_pp": -0.1,
   "field_pnl_net": {
    "A": -473.59,
    "B": -669.03
   },
   "abs_move_delta": {
    "1h": {
     "A": 0.401,
     "B": 0.331,
     "delta_pct_rel": -17.5
    },
    "4h": {
     "A": 0.81,
     "B": 0.664,
     "delta_pct_rel": -18.0
    },
    "1d": {
     "A": 2.352,
     "B": 2.102,
     "delta_pct_rel": -10.6
    }
   },
   "btc_vol_ann_pct": {
    "A": 37.9,
    "B": 35.22
   },
   "note": "descriptive, two adjacent calendar weeks; hit rule = methodology v1.1 trade-based; raw_sign = price-sign robustness check (closes at :00 vs slots at :01)"
  },
  "lineage_note": "lineage splice ACTIVE: grok-4.5 rows are aggregated into grok-4.6 (flip 2026-08-24T09:22:00+00:00, inside week A). Week A (2026-08-24..2026-08-30) mixes the 4.5-era (Aug 24 00:00-09:22) with 4.6 and reproduces the published issue #4; week B is pure grok-4.6."
 },
 "why_it_matters": "MarketMania 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. Issue #5's grok row is a pure grok-4.6 week while issue #4's was a lineage splice, 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 inside the PREVIOUS window (week A, issue #4): every issue-#4 grok number is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\"; this window is pure grok-4.6, so the grok week-over-week row compares a pure-4.6 week with a spliced week.",
 "practical_implications": [
  "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. Five issues, and the best ticker has changed hands every time (SOL -> BNB -> BTC -> SOL -> BNB).",
  "Self-agreement lifts run -17.4pp to -2.5pp with small disagree cells; carry the regime split into the monthly test, do not act on it."
 ],
 "limitations": [
  "Single week (Aug 31-Sep 6, 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 #4: 3 of 7); every comparison with issue #4 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. 10 events across five issues cannot characterise detector performance either way.",
  "Cross-TF and cross-FH disagreement cells are small (n=10, 12, 20, 12); 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. Series density and lineage notes (wave 5): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 31-Sep 6) daily coverage is FULL, as it was in issue #4 (the first full week of the series): the 1w series has slots on 7 of 7 days and the 1M series on 7 of 7 days; both sit outside this report's FH gate (1h/4h/1d) and appear only in the exclusions counter (1w 490, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, inside the PREVIOUS window (week A, issue #4): this window carries 1,470 grok rows and none from the 4.5 era, so every grok number in this issue is a pure grok-4.6 line, while every issue-#4 grok number is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\" -- the grok week-over-week row compares a pure-4.6 week with a spliced week. (3) qwen-3.8-max succeeded qwen-3.7-max on Aug 5, 2026 (fully before this window).",
  "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #4 after the audit that found the raw table mixes two exchanges; issue #4 values are as published.",
  "Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional."
 ],
 "notes": {
  "series_density_and_lineage_wave5": "Series density and lineage notes (wave 5): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 31-Sep 6) daily coverage is FULL, as it was in issue #4 (the first full week of the series): the 1w series has slots on 7 of 7 days and the 1M series on 7 of 7 days; both sit outside this report's FH gate (1h/4h/1d) and appear only in the exclusions counter (1w 490, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, inside the PREVIOUS window (week A, issue #4): this window carries 1,470 grok rows and none from the 4.5 era, so every grok number in this issue is a pure grok-4.6 line, while every issue-#4 grok number is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\" -- the grok week-over-week row compares a pure-4.6 week with a spliced week. (3) qwen-3.8-max succeeded qwen-3.7-max on Aug 5, 2026 (fully before this window).",
  "grok_era": {
   "ids_in_window": [
    "grok-4.6"
   ],
   "n_grok45": 0,
   "n_grok46": 1470,
   "n_grok45_rows": 0,
   "n_grok46_rows": 1470,
   "first_grok46_slot": "2026-08-31T00:01:00+00:00",
   "last_grok45_slot": null,
   "first_grok45_slot": null,
   "flip_at": "2026-08-24T09:22:00+00:00",
   "flip_inside_window": false,
   "prev_window": {
    "n_grok45_rows": 110,
    "n_grok46_rows": 1360
   },
   "merged_model_id": "grok-4.6",
   "merged_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
   "lineage_ids": [
    "grok-4.5",
    "grok-4.6"
   ],
   "lineage_source": "benchmarks.config.lineage_ids('grok-4.6') + MODEL_SUCCESSION[grok-4.6]=('grok-4.5',)",
   "lineage_warnings": [],
   "rows_remapped_a_plus_b": 110,
   "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- inside the PREVIOUS window (week A); all per-model aggregates use the lineage splice grok-4.5 -> grok-4.6 (one row, labelled 'grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)'). Week B (this window) is pure grok-4.6; week A (prev window) mixes both eras and reproduces issue #4.",
   "cells_with_two_grok_rows": 0
  },
  "model_ids_line": [
   "claude-fable-5",
   "claude-opus-5",
   "deepseek-v4-pro",
   "gemini-3.1-pro",
   "gpt-5.6-sol",
   "grok-4.6",
   "qwen-3.8-max"
  ],
  "model_roster": {
   "current": [
    "claude-fable-5",
    "claude-opus-5",
    "deepseek-v4-pro",
    "gemini-3.1-pro",
    "gpt-5.6-sol",
    "grok-4.6",
    "qwen-3.8-max"
   ],
   "current_count": 7,
   "current_raw_ids_in_window": [
    "claude-fable-5",
    "claude-opus-5",
    "deepseek-v4-pro",
    "gemini-3.1-pro",
    "gpt-5.6-sol",
    "grok-4.6",
    "qwen-3.8-max"
   ],
   "archived_legacy": [
    "grok-4.5",
    "qwen-3.7-max",
    "claude-opus-4.8"
   ],
   "archived_legacy_count": 3,
   "archived_legacy_detail": [
    {
     "model": "grok-4.5",
     "n_rows_since": 11128,
     "first_slot": "2026-07-15T18:01:00+00:00",
     "last_slot": "2026-08-24T09:01:00+00:00"
    },
    {
     "model": "qwen-3.7-max",
     "n_rows_since": 7479,
     "first_slot": "2026-07-15T18:01:00+00:00",
     "last_slot": "2026-08-05T09:01:00+00:00"
    },
    {
     "model": "claude-opus-4.8",
     "n_rows_since": 6315,
     "first_slot": "2026-07-15T18:01:00+00:00",
     "last_slot": "2026-07-30T08:01:00+00:00"
    }
   ],
   "archived_excluding_lineage_merged": [
    "qwen-3.7-max",
    "claude-opus-4.8"
   ],
   "tracked_total": 10,
   "canon_expected": {
    "current": 7,
    "archived_legacy": 4,
    "tracked_total": 11
   },
   "matches_canon_11_7_4": false,
   "roster_since": "2026-07-11T00:00:00+00:00",
   "note": "current = merged-lineage model ids emitting inside the window (grok-4.5 rows are spliced into grok-4.6); archived legacy = raw ids seen since 2026-07-11 that no longer emit in the window -- grok-4.5 is one of them by raw id even though its rows are spliced. Models retired before 2026-07-11 are not visible to this query."
  },
  "series_density": {
   "1d": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": true,
    "n_slots": 7,
    "n_rows": 490,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "1h": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": true,
    "n_slots": 168,
    "n_rows": 5880,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "1m": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": false,
    "n_slots": 7,
    "n_rows": 490,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "1w": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": false,
    "n_slots": 7,
    "n_rows": 490,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   },
   "4h": {
    "days_with_slots": 7,
    "expected_days": 7,
    "full_daily_coverage": true,
    "in_fh_gate": true,
    "n_slots": 42,
    "n_rows": 2940,
    "days": [
     "2026-08-31",
     "2026-09-01",
     "2026-09-02",
     "2026-09-03",
     "2026-09-04",
     "2026-09-05",
     "2026-09-06"
    ],
    "missing_days": []
   }
  },
  "week_over_week_note": "Week-over-week columns compare back-to-back windows: Aug 24-30 (issue #4, and week A of this issue's alive slice) vs Aug 31-Sep 6 (this issue). \"Was\" values are the numbers published in issue #4; deltas are computed on unrounded rates.",
  "research_to_date_pin": "Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional.",
  "market_state_audit": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #4 after the audit that found the raw table mixes two exchanges; issue #4 values are as published.",
  "engine": "Engine 1.1 has powered the sandbox since Aug 18, i.e. before this window; no cross-engine PnL comparisons are claimed.",
  "report_template": "PDF built with the canonical mm_report.py template module; preview card with report_preview.render_preview"
 },
 "living_series_note": "**Issue #5.** Weekly Model Watch is a living series. Issue #4's open questions -- does any model hold the top rank two issues running and does the title rule ever fire, and does the reversal detector find callers outside ETH? -- read this issue as: gemini-3.1-pro at the top (46.5%) where issue #4 had deepseek-v4-pro, no weekly title was awarded, for the fifth issue running, and 0 of 2 qualifying turns found a caller, on turns in ETH only. The grok row is pure grok-4.6 this issue; issue #4's grok row was the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\". Engine 1.1 has powered the sandbox since Aug 18, i.e. before this window; no cross-engine PnL comparisons are claimed.",
 "testing_next": [
  "Next issue: does any model hold the top rank two issues running, and does the title rule ever fire?",
  "Does the reversal detector find callers at all, and on a non-ETH turn?",
  "Monthly series: cross-TF/FH lifts with regime control across mixed weeks (trend vs. flat), block bootstrap."
 ],
 "related_research": [
  {
   "title": "Consensus Watch #5",
   "url": "https://marketmania.ai/research/reports/consensus-watch-2026-08-31.pdf"
  },
  {
   "title": "Weekly Calibration #5",
   "url": "https://marketmania.ai/research/reports/weekly-calibration-2026-08-31.pdf"
  },
  {
   "title": "Weekly Model Watch #4",
   "url": "https://marketmania.ai/research/reports/model-watch-2026-08-24.pdf"
  }
 ],
 "related_research_note": "The three weekly reports publish together as one issue each week; each links straight to the others' PDF and to its own previous issue. Direct links are the posting rule from wave 2 on.",
 "research_to_date": {
  "this_report": {
   "scored_observations": 4520
  },
  "platform": {
   "as_of_cutoff": "2026-09-07",
   "resolved_forecasts": 42905,
   "resolved_forecasts_definition": "directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, at the issue cutoff",
   "resolved_forecasts_since": "2026-07-11",
   "models_tracked": 10,
   "models_tracked_detail": "7 models tracked (current line-up; earlier versions folded into their successors' lineage)",
   "assets": 5,
   "forecast_horizons": 5,
   "cadence": "hourly",
   "published_research_reports": 25
  },
  "source": "weekly_metrics_2026-08-31.json",
  "control_reproduces_wave3_pin": true,
  "pin_note": "Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional."
 },
 "citation": {
  "bibtex_key": "mm_modelwatch_2026w36",
  "title": "Weekly Model Watch #5: weekly leaderboard, ticker/pair reads and cross-confirmation, Aug 31-Sep 6 2026",
  "author": "MarketMania Research",
  "year": 2026,
  "month": "September",
  "day": 8,
  "url": "https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf",
  "note": "Methodology v1.1; window Aug 31-Sep 6, 2026 UTC; source weekly_metrics_2026-08-31.json"
 },
 "site_entry": {
  "slug": "model-watch-2026-08-31",
  "series": "weekly",
  "seriesLabel": "Weekly · Model Watch",
  "date": "September 8, 2026",
  "title": "Weekly Model Watch #5",
  "summary": "4,520 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 46.5% (+2.3pp w/w) against deepseek-v4-pro at 44.2%, a gap of 2.3pp. BNB was the most readable ticker at 47.3% and SOL the hardest at 38.7%, with 0 of 5 assets above 51 percent. Reversals: 0 of 2 qualifying turns found a caller. Field accuracy moved 42.9% -> 42.8% (-0.1pp) with 1 of 7 models improving.",
  "stats": [
   "4,520 directional calls",
   "No weekly title again",
   "field -0.1pp w/w"
  ],
  "manifest_key_finding": "Field accuracy moved 42.9% -> 42.8% (-0.1pp) with 1 of 7 models improving; gemini-3.1-pro led the week at 46.5% (+2.3pp w/w) and no weekly title was awarded, for the fifth issue running.",
  "files": {
   "pdf": "/research/reports/model-watch-2026-08-31.pdf",
   "md": "/research/reports/model-watch-2026-08-31.md",
   "json": "/research/reports/model-watch-2026-08-31.json"
  }
 },
 "social": {
  "telegram": {
   "photo_caption_html": "📊 <b>Weekly Model Watch #5</b> — weekly series (window Aug 31–Sep 6)\n\nWho led, what was readable, who saw the turn, does self-agreement help? 4,520 directional calls, 7 model lines:\n• <b>No weekly title again</b> — gemini-3.1-pro leads at <b>46.5%</b> (+2.3pp w/w) against deepseek-v4-pro at 44.2%, a 2.3pp gap\n• Tickers: <b>BNB</b> most readable (47.3%); SOL hardest at 38.7% — 0 of 5 assets above 51%\n• Reversals: 0 of 2 qualifying turns found a caller\n• Market check: field accuracy 42.9% -> 42.8% (-0.1pp), 1 of 7 improved; |1d| move 2.35% -> 2.10%\n\nNote: grok is a pure 4.6 line this window; issue #4's grok row was the 4.5+4.6 splice (flip Aug 24 09:22 UTC)\n\n📄 <a href=\"https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf\">Full report (PDF)</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #models #research",
   "document_caption_html": "📄 Weekly Model Watch #5 — full report (PDF).\nWeb copy: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf\">model-watch-2026-08-31.pdf</a> · machine-readable: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-08-31.json\">JSON</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #models #research",
   "caption_len": 886,
   "document_caption_len": 390,
   "limit": 1024
  },
  "x": {
   "main": "Weekly Model Watch #5 is out.\n\n4,520 directional LLM calls: field accuracy 42.9% -> 42.8% w/w, 1 of 7 models improved. gemini-3.1-pro led at 46.5%; BNB was the most readable ticker.\n\nFull report: https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf\n\n#AI #Crypto #Research #LLM",
   "main_tco_len": 247,
   "reply": "Hit-rate, w/w:\ngemini-3.1: 44.2 -> 46.5\ndeepseek: 44.6 -> 44.2\nfable-5: 43.0 -> 42.8\nqwen-3.8: 42.6 -> 42.5\nopus-5: 42.4 -> 42.4\ngpt-5.6: 42.7 -> 41.6\ngrok-4.6: 39.9 -> 37.6\nField: 42.9 -> 42.8. Wilson CIs; methodology v1.1.\nData: https://marketmania.ai/research/reports/model-watch-2026-08-31.json",
   "reply_tco_len": 254,
   "limit_tco": 280,
   "replies_max": 1
  }
 }
}