{"direction":"long","leverage":750,"take_profit_bp":10,"take_profit_price":81579.00,"bust_distance_bp":8.53,"bust_price_estimate":81427.96,"expected_net_return_on_margin_pct":54.3,"confidence":0.52,"regime":"ranging","key_levels":{"support":81483,"resistance":81563},"entry_mid":81497.5,"data_time_utc":"2026-10-17T04:59:08.140Z","reasoning":"Facts from the Hyperliquid API at 04:59 UTC: the perp mid (81497.5) sits 4.5 bp below oraclePx 81534 (premium -3.75 bp), 5.4 bp below the 6h VWAP and just above the 1h low of 81483, and the book within 10 bp of mid holds 85.3 BTC of bids against 37.0 BTC of asks (17.7 BTC at the best bid), so my inference is a mean-reversion bounce toward the 1h high at 81563 and the oracle. 1m realized vol is only about 1.5 bp per minute (the 1h range is about 10 bp), so I chose the lowest allowed leverage, 750x, because it gives the widest bust buffer at 8.5 bp. I set a 10 bp target because it keeps about 72% of the gain after the haircut (54% on margin), and in a backtest of 30-minute windows on the 6h of 5m candles, longs at 750x hit this target 18 times and busted 21 times, with 33 windows not resolving. Uncertainty: the 24h trend is down (-78 bp, and on both 1m and 15m the 9-period EMA is below the 21-period), and the edge is small.","risks":"The 24h downtrend resuming, or a thin-book weekend sweep below the 81483 low toward the 6h low of 81387, would cross the 81428 bust. The 30 minutes may also end without either level being hit, because 1m volatility is very low."}