{
  "market": "BTC",
  "direction": "short",
  "leverage": 850,
  "take_profit_bp": 10,
  "take_profit_price": 81930.4875,
  "bust_distance_bp": 6.964705882352942,
  "bust_price_estimate": 82069.61929411764,
  "expected_net_return_on_margin_pct": 61.56965106726634,
  "confidence": 0.52,
  "regime": "trending_down",
  "key_levels": {
    "support": 81929,
    "resistance": 82066
  },
  "entry_mid": 82012.5,
  "data_time_utc": "2026-10-19T23:12:39.442629+00:00",
  "reasoning": "Hyperliquid snapshots (https://api.hyperliquid.xyz/info; saved in research/snapshot.json) favor BTC over ETH: BTC's 1m and 5m EMA9 are below EMA21, while ETH's 5m averages still favor an uptrend, and 15m-window minute-return volatility is 2.49 versus 3.44 bp. BTC trades below its falling short-term averages and hourly VWAP, suggesting downward continuation, although the rising 15m EMA and slightly positive sampled trade delta weaken that inference. At 850x, the 6.96 bp bust lies just beyond the recent 82066 rebound high; the 10 bp target sits immediately above observed 81929 support and keeps 61.57% margin profit after the haircut and win fee, conditional on hitting it. Conservative candle replay across 30 overlapping, fully observed 30-minute windows produced 14 targets, seven busts and nine unresolved cases for this BTC setup versus 14 targets and 14 busts for ETH, while a 15 bp BTC target hit only four times; confidence is subjective, not a calibrated success probability.",
  "risks": "A bounce from nearer support at 81978, amplified by the rising 15m trend, could lift the mid through the estimated bust and forfeit the entire margin. The slightly bid-heavy shallow book, sparse trade sample and overlapping trade-price candles cannot establish executable mid-price barrier probabilities or guarantee closure within 30 minutes."
}