{"direction":"long","leverage":750,"take_profit_bp":15,"take_profit_price":81626.95,"bust_distance_bp":8.53,"bust_price_estimate":81435.14,"expected_net_return_on_margin_pct":86.8,"confidence":0.38,"regime":"ranging","key_levels":{"support":81400,"resistance":81546},"entry_mid":81504.695,"data_time_utc":"2026-10-17T05:12:48Z","reasoning":"FACTS (Coinbase BTC-USD public API, api.exchange.coinbase.com /candles at 60s/300s/900s and /ticker, bid 81504.69 / ask 81504.70 at 05:12:48Z): after a 24h drop from about 82831 to 80694, price has made higher lows over roughly 2h (81300, 81383, 81386, 81400, 81468) and the latest 15m candle closed near its high at 81513, about 14 bp above the 81400 base; recent 5m ranges are about 4-7 bp and 1m ranges about 1-4 bp, in quiet Saturday Asia-session volume. INFERENCE: the micro-structure leans slightly up inside a 81300-81550 range, so long is chosen, but the edge is small. With no drift, the chance of hitting TP before bust is about bust/(bust+TP), so expected value per trade scales with keep(TP) x (10000 - 4.8 x leverage)/10000; that favors the lowest allowed leverage (750x, 8.5 bp bust = about $70) and a target big enough to keep about 77% after the haircut (15 bp = about $122, or 86.8% on margin), which a 30-minute window can plausibly reach at current volatility. UNCERTAINTY: Hyperliquid's POST-only info API could not be reached from this run, so entry_mid and the levels come from Coinbase spot and not the Hyperliquid perp mid PaperTrade uses (usually a few dollars apart); funding, OI, the L2 book, CVD and the macro calendar were not checked.","risks":"The bust at about 81435 sits just above the repeatedly tested 81400-81416 cluster, so a normal pullback to range support (or a rejection at 81546-81561 resistance) would liquidate before the 81627 target; the live Hyperliquid mid can also differ from the Coinbase price used here, which moves both levels."}