{
  "report_rows": [
    {
      "kpi": "Room nights growth (Y/Y)",
      "prediction": "BEAT",
      "answer": "pred ~+6% (~328M) vs. cons ~+5% (~324M)",
      "confidence": "MEDIUM"
    },
    {
      "kpi": "Gross bookings growth (Y/Y)",
      "prediction": "BEAT",
      "answer": "pred ~+8% (~$50.4B) vs. cons ~+6% (~$49.5B)",
      "confidence": "MEDIUM"
    },
    {
      "kpi": "Adjusted EBITDA",
      "prediction": "BEAT",
      "answer": "pred ~$2.52B (+7%) vs. cons ~$2.48B (+5%)",
      "confidence": "MEDIUM"
    }
  ],
  "guide_rows": [
    {
      "kpi": "Q3 room nights growth",
      "prediction": "BETTER",
      "answer": "guide ~+6% (~342M) vs. cons ~+5% (~339M) (Q3 2026)",
      "confidence": "MEDIUM"
    },
    {
      "kpi": "Q3 revenue",
      "prediction": "BETTER",
      "answer": "guide ~$9.7B (+8%) vs. cons ~$9.6B (+7%) (Q3 2026)",
      "confidence": "LOW"
    },
    {
      "kpi": "FY2026 gross bookings growth",
      "prediction": "BETTER",
      "answer": "guide ~+10% (raise low end toward LDD) vs. cons ~+9% (FY2026)",
      "confidence": "LOW"
    },
    {
      "kpi": "FY2026 adjusted EPS growth",
      "prediction": "BETTER",
      "answer": "guide ~+15% vs. cons ~+13% (FY2026)",
      "confidence": "LOW"
    },
    {
      "kpi": "Middle East conflict drag / H2 recovery",
      "prediction": "BETTER",
      "answer": "actual drag ~2pt vs. guided ~3pt; recovery starting (Q2/H2 2026)",
      "confidence": "MEDIUM"
    }
  ],
  "day1_residual_pct": 3.0,
  "day1_confidence": "MEDIUM",
  "day5_residual_pct": 1.0,
  "day5_path": "FADE",
  "day5_rationale": "BKNG should beat the conservative 2-4% RN / 4-6% GB guide (Q1 pattern, conflict drag landing lighter at ~2pt) and nudge FY low end higher, driving a positive day-1 residual. But the stock already ran ~13% off the late-July low ($177->$201) before easing to $192.74, so much of the good news is pre-positioned and the bar is high. Out-period math caps follow-through: Q3 is the seasonally largest quarter and management will keep the guide conservative (H2 recovery only 'assumed'), FX benefit fades, and FY EPS is only modestly raised, so estimates drift rather than jump. Net: initial pop partially gives back over five days -> fade.",
  "day5_confidence": "LOW"
}