{
 "slug": "model-watch-2026-08-24",
 "series": "weekly",
 "series_label": "WEEKLY · MODEL WATCH",
 "issue": 4,
 "title": "Weekly Model Watch #4",
 "language": "en",
 "window_start": "2026-08-24T00:00:00+00:00",
 "window_end": "2026-08-31T00:00:00+00:00",
 "window_text": "Aug 24-30, 2026 UTC",
 "cutoff": "2026-08-31T16:00:00+00:00",
 "generated_at": "2026-09-02T12:15:22+00:00",
 "methodology": "v1.1 (2026-08-10)",
 "methodology_hash": "e66c7e8c864a2233",
 "source": "weekly_metrics_2026-08-24.json",
 "hit_rule": "direction: exit tp1/tp2 -> hit, sl -> miss, expiry -> sign of gross pnl",
 "bibtex_key": "mm_modelwatch_2026w35",
 "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": 4576,
  "hit_rate": 0.4294
 },
 "market_state": {
  "panel": "market_state_6p2",
  "week": {
   "start": "2026-08-24",
   "end": "2026-08-31"
  },
  "prev_week": {
   "start": "2026-08-17",
   "end": "2026-08-24"
  },
  "current": {
   "btc_net_pct": -0.07,
   "realized_vol_ann_pct": 30.85,
   "volume_top5_usd_bn": 19.7224,
   "volume_top5_wow_pct": -24.1,
   "avg_pairwise_corr": 0.85
  },
  "previous_as_published_issue3": {
   "btc_net_pct": 23.58,
   "realized_vol_ann_pct": 71.1,
   "volume_top5_usd_bn": 25.97,
   "avg_pairwise_corr": 0.81
  },
  "snapshot_line": "BTC net -0.07% (prior +23.58%) - ann vol 30.9% (was 71.1%) - TOP5 volume $19.72B, -24% w/w - pairwise corr 0.85 (was 0.81)",
  "exchange": "binance",
  "audit_filter_applied": true,
  "source": "single-exchange (binance) daily candle recompute, market_state_2026-08-24.json; prior-week values as published in issue #3",
  "audit_note": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #3 after the audit that found the raw table mixes two exchanges; issue #3 values are as published.",
  "prev_week_control": {
   "published_issue3": {
    "btc_net_pct": 23.58,
    "realized_vol_ann_pct": 71.1,
    "volume_top5_usd_bn": 25.97,
    "avg_pairwise_corr": 0.81
   },
   "got": {
    "btc_net_pct": 23.58,
    "realized_vol_ann_pct": 65.02,
    "volume_top5_usd_bn": 25.9715,
    "avg_pairwise_corr": 0.81
   },
   "match": {
    "btc_net_pct": true,
    "realized_vol_ann_pct": false,
    "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": false,
   "note": "previous week of this run == published issue-#3 week; the four Snapshot pins must reproduce"
  },
  "regime_days": {
   "2026-08-24": {
    "move_pct": 1.61,
    "regime": "trend"
   },
   "2026-08-25": {
    "move_pct": -0.58,
    "regime": "flat"
   },
   "2026-08-26": {
    "move_pct": 0.63,
    "regime": "flat"
   },
   "2026-08-27": {
    "move_pct": 1.53,
    "regime": "trend"
   },
   "2026-08-28": {
    "move_pct": -2.91,
    "regime": "trend"
   },
   "2026-08-29": {
    "move_pct": 0.48,
    "regime": "flat"
   },
   "2026-08-30": {
    "move_pct": -0.7,
    "regime": "flat"
   }
  },
  "alive_slice_context": {
   "abs_move_1d_pct": {
    "A": 4.283,
    "B": 2.352
   },
   "btc_realized_vol_ann_pct_hourly": {
    "A": 58.72,
    "B": 37.9
   }
  }
 },
 "key_finding": {
  "label": "KEY FINDING · OBSERVATION (one weekly window)",
  "text": "Field accuracy went 54.0% -> 42.9% (-11.1pp) with 0 of 7 models improving, and no weekly title was awarded, for the fourth issue running.",
  "evidence_level": "observation"
 },
 "tables": {
  "leaderboard": {
   "rows": [
    {
     "model": "deepseek-v4-pro",
     "n": 729,
     "hit_rate": 0.4458,
     "wilson_95": [
      0.4101,
      0.4821
     ],
     "coverage": 0.5481,
     "model_label": "deepseek-v4-pro",
     "insufficient": false
    },
    {
     "model": "gemini-3.1-pro",
     "n": 755,
     "hit_rate": 0.4424,
     "wilson_95": [
      0.4073,
      0.478
     ],
     "coverage": 0.5685,
     "model_label": "gemini-3.1-pro",
     "insufficient": false
    },
    {
     "model": "claude-fable-5",
     "n": 651,
     "hit_rate": 0.4301,
     "wilson_95": [
      0.3926,
      0.4684
     ],
     "coverage": 0.4921,
     "model_label": "claude-fable-5",
     "insufficient": false
    },
    {
     "model": "gpt-5.6-sol",
     "n": 729,
     "hit_rate": 0.4266,
     "wilson_95": [
      0.3912,
      0.4628
     ],
     "coverage": 0.5481,
     "model_label": "gpt-5.6-sol",
     "insufficient": false
    },
    {
     "model": "qwen-3.8-max",
     "n": 672,
     "hit_rate": 0.4256,
     "wilson_95": [
      0.3887,
      0.4633
     ],
     "coverage": 0.5053,
     "model_label": "qwen-3.8-max",
     "insufficient": false
    },
    {
     "model": "claude-opus-5",
     "n": 561,
     "hit_rate": 0.4242,
     "wilson_95": [
      0.384,
      0.4655
     ],
     "coverage": 0.4218,
     "model_label": "claude-opus-5",
     "insufficient": false
    },
    {
     "model": "grok-4.6",
     "n": 479,
     "hit_rate": 0.3987,
     "wilson_95": [
      0.3559,
      0.4432
     ],
     "coverage": 0.3602,
     "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": 0.3,
   "movement_vs_prior_week": [
    {
     "model": "deepseek-v4-pro",
     "hit_rate_issue3_published": 0.5161,
     "hit_rate_issue4": 0.4458,
     "delta_pp": -7.0,
     "rank_issue4": 1
    },
    {
     "model": "gemini-3.1-pro",
     "hit_rate_issue3_published": 0.5414,
     "hit_rate_issue4": 0.4424,
     "delta_pp": -9.9,
     "rank_issue4": 2
    },
    {
     "model": "claude-fable-5",
     "hit_rate_issue3_published": 0.5648,
     "hit_rate_issue4": 0.4301,
     "delta_pp": -13.5,
     "rank_issue4": 3
    },
    {
     "model": "gpt-5.6-sol",
     "hit_rate_issue3_published": 0.5363,
     "hit_rate_issue4": 0.4266,
     "delta_pp": -11.0,
     "rank_issue4": 4
    },
    {
     "model": "qwen-3.8-max",
     "hit_rate_issue3_published": 0.5227,
     "hit_rate_issue4": 0.4256,
     "delta_pp": -9.7,
     "rank_issue4": 5
    },
    {
     "model": "claude-opus-5",
     "hit_rate_issue3_published": 0.5908,
     "hit_rate_issue4": 0.4242,
     "delta_pp": -16.7,
     "rank_issue4": 6
    },
    {
     "model": "grok-4.6",
     "hit_rate_issue3_published": 0.5141,
     "hit_rate_issue4": 0.3987,
     "delta_pp": -11.5,
     "rank_issue4": 7
    }
   ],
   "note": "No weekly title is awarded this issue -- for the fourth time. Title rule: **>=5pp gap AND non-overlapping 95% CI, N>=10** per side. deepseek-v4-pro's lead over gemini-3.1-pro is 0.3pp and their CIs overlap, so neither condition is met."
  },
  "ticker": {
   "rows": [
    {
     "symbol": "SOL",
     "n": 1025,
     "hit_rate": 0.4576,
     "wilson_95": [
      0.4273,
      0.4882
     ],
     "issue3_hit_rate": 0.5434
    },
    {
     "symbol": "BTC",
     "n": 940,
     "hit_rate": 0.4372,
     "wilson_95": [
      0.4058,
      0.4691
     ],
     "issue3_hit_rate": 0.5751
    },
    {
     "symbol": "ETH",
     "n": 827,
     "hit_rate": 0.4329,
     "wilson_95": [
      0.3995,
      0.4669
     ],
     "issue3_hit_rate": 0.5279
    },
    {
     "symbol": "BNB",
     "n": 895,
     "hit_rate": 0.4123,
     "wilson_95": [
      0.3805,
      0.4448
     ],
     "issue3_hit_rate": 0.5397
    },
    {
     "symbol": "XRP",
     "n": 889,
     "hit_rate": 0.4027,
     "wilson_95": [
      0.3709,
      0.4353
     ],
     "issue3_hit_rate": 0.5124
    }
   ],
   "best": "SOL",
   "hardest": "XRP",
   "all_above_51pct": false
  },
  "pairs_top": [
   {
    "model": "qwen-3.8-max",
    "symbol": "SOL",
    "n": 150,
    "hit_rate": 0.4867
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "ETH",
    "n": 130,
    "hit_rate": 0.4769
   },
   {
    "model": "claude-opus-5",
    "symbol": "SOL",
    "n": 131,
    "hit_rate": 0.4733
   },
   {
    "model": "deepseek-v4-pro",
    "symbol": "SOL",
    "n": 154,
    "hit_rate": 0.461
   },
   {
    "model": "gemini-3.1-pro",
    "symbol": "XRP",
    "n": 154,
    "hit_rate": 0.461
   },
   {
    "model": "gpt-5.6-sol",
    "symbol": "SOL",
    "n": 157,
    "hit_rate": 0.4586
   }
  ],
  "pairs_worst": [
   {
    "model": "grok-4.6",
    "symbol": "ETH",
    "n": 81,
    "hit_rate": 0.3827
   },
   {
    "model": "claude-fable-5",
    "symbol": "XRP",
    "n": 116,
    "hit_rate": 0.3707
   },
   {
    "model": "qwen-3.8-max",
    "symbol": "XRP",
    "n": 135,
    "hit_rate": 0.363
   },
   {
    "model": "grok-4.6",
    "symbol": "BNB",
    "n": 100,
    "hit_rate": 0.35
   },
   {
    "model": "claude-opus-5",
    "symbol": "XRP",
    "n": 104,
    "hit_rate": 0.3462
   }
  ],
  "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": "2026-08-28T12:00:00+00:00",
     "new_side": "short",
     "move_8h_pct": -2.49,
     "n_callers": 8,
     "first_caller": {
      "model": "gpt-5.6-sol",
      "slot": "2026-08-28T04:01:00+00:00",
      "fh": "4h",
      "tf": "1h",
      "confidence": 62
     },
     "caller_models": [
      "deepseek-v4-pro",
      "gpt-5.6-sol",
      "grok-4.6",
      "qwen-3.8-max"
     ]
    },
    {
     "symbol": "ETH",
     "t": "2026-08-26T16:00:00+00:00",
     "new_side": "long",
     "move_8h_pct": 2.31,
     "n_callers": 13,
     "first_caller": {
      "model": "gpt-5.6-sol",
      "slot": "2026-08-27T00:01:00+00:00",
      "fh": "1d",
      "tf": "4h",
      "confidence": 69
     },
     "caller_models": [
      "claude-fable-5",
      "claude-opus-5",
      "deepseek-v4-pro",
      "gemini-3.1-pro",
      "gpt-5.6-sol",
      "grok-4.6",
      "qwen-3.8-max"
     ]
    },
    {
     "symbol": "ETH",
     "t": "2026-08-28T12:00:00+00:00",
     "new_side": "short",
     "move_8h_pct": -2.8,
     "n_callers": 0,
     "first_caller": null,
     "caller_models": []
    },
    {
     "symbol": "ETH",
     "t": "2026-08-30T16:00:00+00:00",
     "new_side": "short",
     "move_8h_pct": -2.52,
     "n_callers": 1,
     "first_caller": {
      "model": "deepseek-v4-pro",
      "slot": "2026-08-30T04:01:00+00:00",
      "fh": "4h",
      "tf": "4h",
      "confidence": 56
     },
     "caller_models": [
      "deepseek-v4-pro"
     ]
    }
   ],
   "scope_v1": "BTC/ETH only",
   "status": "ok"
  },
  "cross_tf": {
   "fh_4h_tf_4h_vs_1h": {
    "agree": {
     "n": 473,
     "hit_rate": 0.3911,
     "wilson_95": [
      0.3482,
      0.4358
     ],
     "by_regime": {
      "trend": {
       "n": 248,
       "hit_rate": 0.4194
      },
      "flat": {
       "n": 225,
       "hit_rate": 0.36
      }
     }
    },
    "disagree": {
     "n": 18,
     "hit_rate": 0.1111,
     "wilson_95": [
      0.031,
      0.328
     ],
     "by_regime": {
      "trend": {
       "n": 9,
       "hit_rate": 0.2222
      },
      "flat": {
       "n": 9,
       "hit_rate": 0.0
      }
     }
    },
    "mixed_pairs": 218,
    "base": {
     "n": 712,
     "hit_rate": 0.4101
    },
    "lift_pp": -1.9
   },
   "fh_1d_tf_1d_vs_4h": {
    "agree": {
     "n": 105,
     "hit_rate": 0.4952,
     "wilson_95": [
      0.4015,
      0.5893
     ],
     "by_regime": {
      "trend": {
       "n": 71,
       "hit_rate": 0.6056
      },
      "flat": {
       "n": 34,
       "hit_rate": 0.2647
      }
     }
    },
    "disagree": {
     "n": 10,
     "hit_rate": 1.0,
     "wilson_95": [
      0.7225,
      1.0
     ],
     "by_regime": {
      "trend": {
       "n": 1,
       "hit_rate": 1.0
      },
      "flat": {
       "n": 9,
       "hit_rate": 1.0
      }
     }
    },
    "mixed_pairs": 28,
    "base": {
     "n": 143,
     "hit_rate": 0.5594
    },
    "lift_pp": -6.4
   }
  },
  "cross_fh": {
   "fh_4h_confirmed_by_1h": {
    "agree": {
     "n": 429,
     "hit_rate": 0.3893,
     "wilson_95": [
      0.3443,
      0.4362
     ],
     "by_regime": {
      "trend": {
       "n": 234,
       "hit_rate": 0.4145
      },
      "flat": {
       "n": 195,
       "hit_rate": 0.359
      }
     }
    },
    "disagree": {
     "n": 25,
     "hit_rate": 0.24,
     "wilson_95": [
      0.115,
      0.4343
     ],
     "by_regime": {
      "trend": {
       "n": 15,
       "hit_rate": 0.3333
      },
      "flat": {
       "n": 10,
       "hit_rate": 0.1
      }
     }
    },
    "mixed_pairs": 258,
    "base": {
     "n": 712,
     "hit_rate": 0.4101
    },
    "lift_pp": -2.1
   },
   "fh_1d_confirmed_by_4h": {
    "agree": {
     "n": 89,
     "hit_rate": 0.4831,
     "wilson_95": [
      0.3822,
      0.5855
     ],
     "by_regime": {
      "trend": {
       "n": 59,
       "hit_rate": 0.6271
      },
      "flat": {
       "n": 30,
       "hit_rate": 0.2
      }
     }
    },
    "disagree": {
     "n": 10,
     "hit_rate": 0.9,
     "wilson_95": [
      0.5958,
      0.9821
     ],
     "by_regime": {
      "trend": {
       "n": 1,
       "hit_rate": 1.0
      },
      "flat": {
       "n": 9,
       "hit_rate": 0.8889
      }
     }
    },
    "mixed_pairs": 44,
    "base": {
     "n": 143,
     "hit_rate": 0.5594
    },
    "lift_pp": -7.6
   }
  },
  "regime_days": {
   "2026-08-24": {
    "move_pct": 1.61,
    "regime": "trend"
   },
   "2026-08-25": {
    "move_pct": -0.58,
    "regime": "flat"
   },
   "2026-08-26": {
    "move_pct": 0.63,
    "regime": "flat"
   },
   "2026-08-27": {
    "move_pct": 1.53,
    "regime": "trend"
   },
   "2026-08-28": {
    "move_pct": -2.91,
    "regime": "trend"
   },
   "2026-08-29": {
    "move_pct": 0.48,
    "regime": "flat"
   },
   "2026-08-30": {
    "move_pct": -0.7,
    "regime": "flat"
   }
  },
  "side_mix": [
   {
    "model": "claude-fable-5",
    "n": 1323,
    "long": 0.3477,
    "short": 0.1444,
    "sideways": 0.5079,
    "ls_ratio": 2.41
   },
   {
    "model": "claude-opus-5",
    "n": 1330,
    "long": 0.2917,
    "short": 0.1301,
    "sideways": 0.5782,
    "ls_ratio": 2.24
   },
   {
    "model": "deepseek-v4-pro",
    "n": 1330,
    "long": 0.3233,
    "short": 0.2248,
    "sideways": 0.4519,
    "ls_ratio": 1.44
   },
   {
    "model": "gemini-3.1-pro",
    "n": 1328,
    "long": 0.3524,
    "short": 0.2161,
    "sideways": 0.4315,
    "ls_ratio": 1.63
   },
   {
    "model": "gpt-5.6-sol",
    "n": 1330,
    "long": 0.303,
    "short": 0.2451,
    "sideways": 0.4519,
    "ls_ratio": 1.24
   },
   {
    "model": "grok-4.6",
    "n": 1330,
    "long": 0.1827,
    "short": 0.1774,
    "sideways": 0.6398,
    "ls_ratio": 1.03
   },
   {
    "model": "qwen-3.8-max",
    "n": 1330,
    "long": 0.3158,
    "short": 0.1895,
    "sideways": 0.4947,
    "ls_ratio": 1.67
   }
  ],
  "long_short_ratio_field": {
   "value": 1.66,
   "method": "mean of the per-model long-share / short-share ratios"
  },
  "consensus_skew": {
   "cells_ge5_same_side": 539,
   "n_mature": 3371,
   "hit_rate": 0.3996
  },
  "fail_of_week": {
   "model": "gemini-3.1-pro",
   "symbol": "SOL",
   "side": "long",
   "confidence": 85,
   "slot": "2026-08-24T16:01:00+00:00",
   "fh": "4h",
   "tf": "4h",
   "exit_reason": "expiry",
   "pnl_net_usd": -0.28
  },
  "counters": {
   "forecasts_total": 10290,
   "ok_in_gate": 9301,
   "out_of_gate": 978,
   "out_of_gate_by_fh": {
    "1w": 488,
    "1m": 490
   },
   "invalid": 11,
   "mature": 9301,
   "mature_directional": 4576,
   "mature_sideways": 4725,
   "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": {
    "2026-08-24": {
     "move_pct": 1.61,
     "regime": "trend"
    },
    "2026-08-25": {
     "move_pct": -0.58,
     "regime": "flat"
    },
    "2026-08-26": {
     "move_pct": 0.63,
     "regime": "flat"
    },
    "2026-08-27": {
     "move_pct": 1.53,
     "regime": "trend"
    },
    "2026-08-28": {
     "move_pct": -2.91,
     "regime": "trend"
    },
    "2026-08-29": {
     "move_pct": 0.48,
     "regime": "flat"
    },
    "2026-08-30": {
     "move_pct": -0.7,
     "regime": "flat"
    }
   },
   "grok_era": {
    "ids_in_window": [
     "grok-4.5",
     "grok-4.6"
    ],
    "n_grok45": 110,
    "n_grok46": 1360,
    "n_grok45_rows": 110,
    "n_grok46_rows": 1360,
    "first_grok46_slot": "2026-08-24T10:01:00+00:00",
    "last_grok45_slot": "2026-08-24T09:01:00+00:00",
    "first_grok45_slot": "2026-08-24T00:01:00+00:00",
    "flip_at": "2026-08-24T09:22:00+00:00",
    "flip_inside_window": true,
    "prev_window": {
     "n_grok45_rows": 1370,
     "n_grok46_rows": 0
    },
    "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": 1480,
    "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- INSIDE this 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)'). Week A (prev window) is pure grok-4.5 and its numbers reproduce issue #3.",
    "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": [
      "2026-08-24",
      "2026-08-25",
      "2026-08-26",
      "2026-08-27",
      "2026-08-28",
      "2026-08-29",
      "2026-08-30"
     ],
     "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": [
      "2026-08-24",
      "2026-08-25",
      "2026-08-26",
      "2026-08-27",
      "2026-08-28",
      "2026-08-29",
      "2026-08-30"
     ],
     "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": [
      "2026-08-24",
      "2026-08-25",
      "2026-08-26",
      "2026-08-27",
      "2026-08-28",
      "2026-08-29",
      "2026-08-30"
     ],
     "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": [
      "2026-08-24",
      "2026-08-25",
      "2026-08-26",
      "2026-08-27",
      "2026-08-28",
      "2026-08-29",
      "2026-08-30"
     ],
     "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": [
      "2026-08-24",
      "2026-08-25",
      "2026-08-26",
      "2026-08-27",
      "2026-08-28",
      "2026-08-29",
      "2026-08-30"
     ],
     "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-4.5",
     "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": {
   "deepseek-v4-pro": 1.0,
   "qwen-3.8-max": 0.9986,
   "claude-fable-5": 0.9952,
   "claude-opus-5": 1.0,
   "grok-4.6": 1.0,
   "gemini-3.1-pro": 0.9986,
   "gpt-5.6-sol": 1.0
  },
  "calibration_bridge": {
   "best_calibrated_model": "deepseek-v4-pro",
   "brier": 0.276,
   "hit_rate_leader": "deepseek-v4-pro"
  }
 },
 "key_findings": [
  "**OBSERVATION -- No weekly title again.** deepseek-v4-pro leads descriptively at 44.6% [41.0%, 48.2%]; the gap to #2 (gemini-3.1-pro, 44.2%) is 0.3pp. 0 of 7 models improved week over week; issue #3's leader claude-opus-5 is now 6 of 7.",
  "**No ticker cleared 51%.** SOL is the week's most readable coin (45.8%, n=1,025) and XRP the hardest at 40.3%; 0 of the 5 assets cleared 51%. Pair of the week: qwen-3.8-max x SOL 48.7% (n=150); worst: claude-opus-5 x XRP 34.6% (n=104).",
  "**3 of 4 qualifying turns found a caller.** BTC short -2.49%, 8 caller(s), first gpt-5.6-sol; ETH long +2.31%, 13 caller(s), first gpt-5.6-sol; ETH short -2.80%, 0 caller(s), first NOBODY; ETH short -2.52%, 1 caller(s), first deepseek-v4-pro.",
  "**Self-agreement read as a penalty in every cell.** Cross-TF/FH lifts ran -7.6pp to -1.9pp (issue #3: -0.6 to +4.6pp). On trend days, 4h TF-agreement hit 41.9% (n=248) against 36.0% (n=225) on flat days."
 ],
 "market_check": {
  "label": "MARKET CHECK · OBSERVATION (two adjacent calendar weeks)",
  "text": "OBSERVATION — mean |1d move| 4.28% -> 2.35% (-45% rel), BTC realized vol 58.7% -> 37.9% (ann., hourly); field directional accuracy 54.0% -> 42.9% (-11.1pp), 0 of 7 models improved, sim win-rate up for 0 of 7, field sim PnL +$889.37 -> -$473.59 (adjacent calendar weeks Aug 17-23 vs Aug 24-30; descriptive, one pair of weeks, not a claim).",
  "robustness": "Robustness: raw price-sign accuracy 53.5% -> 44.2% — the same direction as the trade-based hit rule.",
  "weeks": {
   "A": {
    "days": "2026-08-17..2026-08-23",
    "slots": [
     "2026-08-17T00:00:00+00:00",
     "2026-08-24T00:00:00+00:00"
    ],
    "cutoff": "2026-08-24T16:00:00+00:00"
   },
   "B": {
    "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"
   }
  },
  "price_source": "crypto_spot/1h@:00",
  "accuracy_field": {
   "A": {
    "n": 5130,
    "hit_rate": 0.5405,
    "wilson_95": [
     0.5269,
     0.5541
    ]
   },
   "B": {
    "n": 4576,
    "hit_rate": 0.4294,
    "wilson_95": [
     0.4151,
     0.4438
    ]
   },
   "delta_pp": -11.1
  },
  "accuracy_per_model": [
   {
    "model": "deepseek-v4-pro",
    "model_label": "deepseek-v4-pro",
    "A": {
     "n": 777,
     "hit_rate": 0.5161,
     "wilson_95": [
      0.481,
      0.5511
     ]
    },
    "B": {
     "n": 729,
     "hit_rate": 0.4458,
     "wilson_95": [
      0.4101,
      0.4821
     ]
    },
    "delta_pp": -7.0
   },
   {
    "model": "qwen-3.8-max",
    "model_label": "qwen-3.8-max",
    "A": {
     "n": 706,
     "hit_rate": 0.5227,
     "wilson_95": [
      0.4858,
      0.5593
     ]
    },
    "B": {
     "n": 672,
     "hit_rate": 0.4256,
     "wilson_95": [
      0.3887,
      0.4633
     ]
    },
    "delta_pp": -9.7
   },
   {
    "model": "gemini-3.1-pro",
    "model_label": "gemini-3.1-pro",
    "A": {
     "n": 798,
     "hit_rate": 0.5414,
     "wilson_95": [
      0.5067,
      0.5756
     ]
    },
    "B": {
     "n": 755,
     "hit_rate": 0.4424,
     "wilson_95": [
      0.4073,
      0.478
     ]
    },
    "delta_pp": -9.9
   },
   {
    "model": "gpt-5.6-sol",
    "model_label": "gpt-5.6-sol",
    "A": {
     "n": 688,
     "hit_rate": 0.5363,
     "wilson_95": [
      0.499,
      0.5733
     ]
    },
    "B": {
     "n": 729,
     "hit_rate": 0.4266,
     "wilson_95": [
      0.3912,
      0.4628
     ]
    },
    "delta_pp": -11.0
   },
   {
    "model": "grok-4.6",
    "model_label": "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22 UTC)",
    "A": {
     "n": 710,
     "hit_rate": 0.5141,
     "wilson_95": [
      0.4773,
      0.5507
     ]
    },
    "B": {
     "n": 479,
     "hit_rate": 0.3987,
     "wilson_95": [
      0.3559,
      0.4432
     ]
    },
    "delta_pp": -11.5
   },
   {
    "model": "claude-fable-5",
    "model_label": "claude-fable-5",
    "A": {
     "n": 779,
     "hit_rate": 0.5648,
     "wilson_95": [
      0.5298,
      0.5992
     ]
    },
    "B": {
     "n": 651,
     "hit_rate": 0.4301,
     "wilson_95": [
      0.3926,
      0.4684
     ]
    },
    "delta_pp": -13.5
   },
   {
    "model": "claude-opus-5",
    "model_label": "claude-opus-5",
    "A": {
     "n": 672,
     "hit_rate": 0.5908,
     "wilson_95": [
      0.5532,
      0.6273
     ]
    },
    "B": {
     "n": 561,
     "hit_rate": 0.4242,
     "wilson_95": [
      0.384,
      0.4655
     ]
    },
    "delta_pp": -16.7
   }
  ],
  "raw_sign_field": {
   "A": {
    "n_scored": 5118,
    "hit_rate": 0.5348
   },
   "B": {
    "n_scored": 4567,
    "hit_rate": 0.4419
   }
  },
  "abs_move_1d": {
   "A": 4.283,
   "B": 2.352,
   "delta_pct_rel": -45.1
  },
  "btc_realized_vol_ann_pct_hourly": {
   "A": 58.72,
   "B": 37.9
  },
  "trading_field": {
   "A": {
    "n_trades": 5130,
    "wr": 0.4626,
    "pnl_net_usd": 889.37,
    "pnl_gross_usd": 1402.37
   },
   "B": {
    "n_trades": 4576,
    "wr": 0.3702,
    "pnl_net_usd": -473.59,
    "pnl_gross_usd": -15.99
   }
  },
  "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": -11.1,
   "field_pnl_net": {
    "A": 889.37,
    "B": -473.59
   },
   "abs_move_delta": {
    "1h": {
     "A": 0.541,
     "B": 0.401,
     "delta_pct_rel": -25.9
    },
    "4h": {
     "A": 1.314,
     "B": 0.81,
     "delta_pct_rel": -38.4
    },
    "1d": {
     "A": 4.283,
     "B": 2.352,
     "delta_pct_rel": -45.1
    }
   },
   "btc_vol_ann_pct": {
    "A": 58.72,
    "B": 37.9
   },
   "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 B). Week A (2026-08-17..2026-08-23) is pure grok-4.5 and reproduces the published issue #3; week B mixes the 4.5-era (Aug 24 00:00-09:22) with 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 #4 is the first issue whose grok row is a lineage splice, so read that 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 row is the lineage splice grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22).",
 "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. Four issues, and the best ticker has changed hands every time (SOL -> BNB -> BTC -> SOL).",
  "Self-agreement lifts run -7.6pp to -1.9pp with small disagree cells; carry the regime split into the monthly test, do not act on it."
 ],
 "limitations": [
  "Single week (Aug 24-30, 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 #3: 5 of 7); every comparison with issue #3 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. 8 events across four issues cannot characterise detector performance either way.",
  "Cross-TF and cross-FH disagreement cells are small (n=18, 10, 25, 10); 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 (spliced), qwen-3.8-max. Series density and lineage notes (wave 4): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 24-30) daily coverage is FULL for the first time in 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 488, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, INSIDE this window: every grok row in this issue is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\" (110 rows from the 4.5 era, 1,360 from 4.6); issue #3 was pure grok-4.5, so the grok week-over-week row joins a spliced line to a single-model line. (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 #3 after the audit that found the raw table mixes two exchanges; issue #3 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 (Aug 31 16:00 UTC) -- the definition pinned in issue #3, which printed 33,809 at the Aug 24 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 38,385 is an additive step under one definition; counter deltas against issue #2 and earlier remain definitional."
 ],
 "notes": {
  "series_density_and_lineage_wave4": "Series density and lineage notes (wave 4): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 24-30) daily coverage is FULL for the first time in 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 488, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, INSIDE this window: every grok row in this issue is the lineage splice \"grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)\" (110 rows from the 4.5 era, 1,360 from 4.6); issue #3 was pure grok-4.5, so the grok week-over-week row joins a spliced line to a single-model line. (3) qwen-3.8-max succeeded qwen-3.7-max on Aug 5, 2026 (fully before this window).",
  "grok_era": {
   "ids_in_window": [
    "grok-4.5",
    "grok-4.6"
   ],
   "n_grok45": 110,
   "n_grok46": 1360,
   "n_grok45_rows": 110,
   "n_grok46_rows": 1360,
   "first_grok46_slot": "2026-08-24T10:01:00+00:00",
   "last_grok45_slot": "2026-08-24T09:01:00+00:00",
   "first_grok45_slot": "2026-08-24T00:01:00+00:00",
   "flip_at": "2026-08-24T09:22:00+00:00",
   "flip_inside_window": true,
   "prev_window": {
    "n_grok45_rows": 1370,
    "n_grok46_rows": 0
   },
   "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": 1480,
   "note": "grok-4.6 went live 2026-08-24T09:22:00+00:00 -- INSIDE this 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)'). Week A (prev window) is pure grok-4.5 and its numbers reproduce issue #3.",
   "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-4.5",
    "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": [
     "2026-08-24",
     "2026-08-25",
     "2026-08-26",
     "2026-08-27",
     "2026-08-28",
     "2026-08-29",
     "2026-08-30"
    ],
    "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": [
     "2026-08-24",
     "2026-08-25",
     "2026-08-26",
     "2026-08-27",
     "2026-08-28",
     "2026-08-29",
     "2026-08-30"
    ],
    "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": [
     "2026-08-24",
     "2026-08-25",
     "2026-08-26",
     "2026-08-27",
     "2026-08-28",
     "2026-08-29",
     "2026-08-30"
    ],
    "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": [
     "2026-08-24",
     "2026-08-25",
     "2026-08-26",
     "2026-08-27",
     "2026-08-28",
     "2026-08-29",
     "2026-08-30"
    ],
    "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": [
     "2026-08-24",
     "2026-08-25",
     "2026-08-26",
     "2026-08-27",
     "2026-08-28",
     "2026-08-29",
     "2026-08-30"
    ],
    "missing_days": []
   }
  },
  "week_over_week_note": "Week-over-week columns compare back-to-back windows: Aug 17-23 (issue #3, and week A of this issue's alive slice) vs Aug 24-30 (this issue). \"Was\" values are the numbers published in issue #3; 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 (Aug 31 16:00 UTC) -- the definition pinned in issue #3, which printed 33,809 at the Aug 24 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 38,385 is an additive step under one definition; counter deltas against issue #2 and earlier remain definitional.",
  "market_state_audit": "Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #3 after the audit that found the raw table mixes two exchanges; issue #3 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 #4.** Weekly Model Watch is a living series. Issue #3's open questions -- does the leaderboard reshuffle again, does the reversal detector find callers on a non-ETH turn, and does the every-ticker-above-51% floor survive? -- read this issue as: deepseek-v4-pro at the top (44.6%), 3 of 4 qualifying turns found a caller, and 0 of 5 tickers above 51%. The grok row is a lineage splice this issue (grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)) because the 4.5 -> 4.6 flip happened inside the window. 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 outside ETH?",
  "Monthly series: cross-TF/FH lifts with regime control across mixed weeks (trend vs. flat), block bootstrap."
 ],
 "related_research": [
  {
   "title": "Consensus Watch #4",
   "url": "https://marketmania.ai/research/reports/consensus-watch-2026-08-24.pdf"
  },
  {
   "title": "Weekly Calibration #4",
   "url": "https://marketmania.ai/research/reports/weekly-calibration-2026-08-24.pdf"
  },
  {
   "title": "Weekly Model Watch #3",
   "url": "https://marketmania.ai/research/reports/model-watch-2026-08-17.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": 4576
  },
  "platform": {
   "as_of_cutoff": "2026-08-31",
   "resolved_forecasts": 38385,
   "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": 17
  },
  "source": "weekly_metrics_2026-08-24.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 (Aug 31 16:00 UTC) -- the definition pinned in issue #3, which printed 33,809 at the Aug 24 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 38,385 is an additive step under one definition; counter deltas against issue #2 and earlier remain definitional."
 },
 "citation": {
  "bibtex_key": "mm_modelwatch_2026w35",
  "title": "Weekly Model Watch #4: weekly leaderboard, ticker/pair reads and cross-confirmation, Aug 24-30 2026",
  "author": "MarketMania Research",
  "year": 2026,
  "month": "September",
  "day": 2,
  "url": "https://marketmania.ai/research/reports/model-watch-2026-08-24.pdf",
  "note": "Methodology v1.1; window Aug 24-30, 2026 UTC; source weekly_metrics_2026-08-24.json"
 },
 "site_entry": {
  "slug": "model-watch-2026-08-24",
  "series": "weekly",
  "seriesLabel": "Weekly · Model Watch",
  "date": "September 2, 2026",
  "title": "Weekly Model Watch #4",
  "summary": "4,576 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 — deepseek-v4-pro led at 44.6% (-7.0pp w/w) against gemini-3.1-pro at 44.2%, a gap of 0.3pp. SOL was the most readable ticker at 45.8% and XRP the hardest at 40.3%, with 0 of 5 assets above 51 percent. Reversals: 3 of 4 qualifying turns found a caller. Field accuracy moved 54.0% -> 42.9% (-11.1pp) with 0 of 7 models improving.",
  "stats": [
   "4,576 directional calls",
   "No weekly title again",
   "field -11.1pp w/w"
  ],
  "manifest_key_finding": "Field accuracy moved 54.0% -> 42.9% (-11.1pp) with 0 of 7 models improving; deepseek-v4-pro led the week at 44.6% (-7.0pp w/w) and no weekly title was awarded, for the fourth issue running.",
  "files": {
   "pdf": "/research/reports/model-watch-2026-08-24.pdf",
   "md": "/research/reports/model-watch-2026-08-24.md",
   "json": "/research/reports/model-watch-2026-08-24.json"
  }
 },
 "social": {
  "telegram": {
   "photo_caption_html": "📊 <b>Weekly Model Watch #4</b> — weekly series (window Aug 24–30)\n\nWho led, what was readable, who saw the turn, does self-agreement help? 4,576 directional calls, 7 model lines:\n• <b>No weekly title again</b> — deepseek-v4-pro leads at <b>44.6%</b> (-7.0pp w/w) against gemini-3.1-pro at 44.2%, a 0.3pp gap\n• Tickers: <b>SOL</b> most readable (45.8%); XRP hardest at 40.3% — 0 of 5 assets above 51%\n• Reversals: 3 of 4 qualifying turns found a caller\n• Market check: field accuracy 54.0% -> 42.9% (-11.1pp), 0 of 7 improved; |1d| move 4.28% -> 2.35%\n\nNote: grok flipped 4.5 -> 4.6 inside this window; grok rows are one spliced line (4.5 era Aug 24 00:00-09:22 + 4.6)\n\n📄 <a href=\"https://marketmania.ai/research/reports/model-watch-2026-08-24.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 #4 — full report (PDF).\nWeb copy: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-08-24.pdf\">model-watch-2026-08-24.pdf</a> · machine-readable: <a href=\"https://marketmania.ai/research/reports/model-watch-2026-08-24.json\">JSON</a> · <a href=\"https://marketmania.ai/research\">all research</a>\n\n#cryptotrading #LLM #AI #crypto #trading #models #research",
   "caption_len": 890,
   "document_caption_len": 390,
   "limit": 1024
  },
  "x": {
   "main": "Weekly Model Watch #4 is out.\n\n4,576 directional LLM calls: field accuracy 54.0% -> 42.9% w/w, 0 of 7 models improved. deepseek-v4-pro led at 44.6%; SOL was the most readable ticker.\n\nFull report: https://marketmania.ai/research/reports/model-watch-2026-08-24.pdf\n\n#AI #Crypto #Research #LLM",
   "main_tco_len": 248,
   "reply": "Hit-rate, w/w:\ndeepseek: 51.6 -> 44.6\ngemini-3.1: 54.1 -> 44.2\nfable-5: 56.5 -> 43.0\ngpt-5.6: 53.6 -> 42.7\nqwen-3.8: 52.3 -> 42.6\nopus-5: 59.1 -> 42.4\ngrok-4.6: 51.4 -> 39.9\nField: 54.0 -> 42.9. Wilson CIs; methodology v1.1.\nData: https://marketmania.ai/research/reports/model-watch-2026-08-24.json",
   "reply_tco_len": 254,
   "limit_tco": 280,
   "replies_max": 1
  }
 }
}