{"direction":"short","leverage":750,"take_profit_bp":10,"take_profit_price":81466.0,"bust_distance_bp":8.53,"bust_price_estimate":81617.1,"expected_net_return_on_margin_pct":55.4,"confidence":0.4,"regime":"trending_down","key_levels":{"support":81525.0,"resistance":81591.0},"entry_mid":81547.5,"data_time_utc":"2026-10-17T11:56:24Z","reasoning":"Facts (Hyperliquid info API, fetched 2026-10-17T11:56:24Z): mid 81547.5 sits below the 1m/5m/15m EMA9 and EMA21 (5m EMA21 81582, 15m EMA21 81599), 24h change is -128.6 bp against prevDayPx 82610, price is 2.9 bp under the 24h VWAP (81571), and the last 1m candle rejected 81591 and closed at the lows; recent prints were sell-dominated (0.225 BTC sold vs 0.007 bought, only 10 trades, so a small sample). In a 30-bar first-passage replay on the last 60 1m candles, a short at 750x (bust 8.53 bp) never busted, while a long busted in 22 of 61 start points; that hour's drift was down, which is my inference for keeping the short bias. Volatility is low on a Saturday (1m realized vol about 1.5-1.7 bp per minute, roughly 9 bp per sqrt(30 min)), so I use the lowest allowed leverage, 750x, for the widest bust buffer and a 10 bp target that is about one 30-minute sigma. That target clears most of the haircut (about 74% kept, about +55% on margin), while the 15-25 bp targets almost never filled within 30 minutes in the replay.","risks":"An 8.5 bp squeeze to about 81617 would liquidate. Two things could cause it: the perp trades at a -4.3 bp premium to oracle (81583), which could pull the mark back up, and the book within 10 bp is bid-heavy (78.5 vs 50.5 BTC). Unanswered: I checked no news, macro calendar or liquidation heatmap, so a weekend headline spike would not be priced into this decision."}