{
  "direction": "long",
  "leverage": 850,
  "take_profit_bp": 12,
  "take_profit_price": 81692.4134,
  "bust_distance_bp": 6.964705882352942,
  "bust_price_estimate": 81537.67183058824,
  "expected_net_return_on_margin_pct": 76.22769094884256,
  "confidence": 0.53,
  "regime": "trending_up",
  "key_levels": {
    "support": 81571,
    "resistance": 81622
  },
  "entry_mid": 81594.5,
  "data_time_utc": "2026-10-17T19:53:15.561436Z",
  "reasoning": "Hyperliquid public API snapshots (https://api.hyperliquid.xyz/info; requests and timestamps in evidence/) show completed-candle 1m/5m/15m EMA9 above EMA21, with the last 30 completed minutes gaining 11.29 bp, favoring a long continuation. Calculations in evidence/metrics.json show 1m open-to-close volatility of 1.94 bp over 15 minutes versus 1.74 bp over 60, and initial visible book bid/ask imbalance of +17.6%, although that snapshot covers only about 2.5 bp per side. At 850x, the 6.96 bp bust allowance balances amplification and pullback room; a 12 bp target requires breaking 81622 but retains about 75% of the gross price gain after the supplied haircut and win fee. In 31 overlapping historical 30-minute long trials, this setting produced 10 targets, 11 busts and 10 unresolved outcomes versus only one target at 16 bp, supporting the smaller target as an inference rather than an established profitable edge.",
  "risks": "Rejection at 81622 or a reversal through 81571 could carry the mid to liquidation and forfeit all margin. The small overlapping OHLC sample is not a calibrated win probability, book liquidity can disappear, and a target within 30 minutes is uncertain."
}