LinkDailyVolumeSurgeBullishBarTrendFilterLong
Hypotheses
LINK Daily Volume Surge Bullish Bar Long-Only with Trend Filter (BINANCE USD-M Futures, Daily, OHLCV-Only)
Hypotheses
Long-only LINK daily strategy on BINANCE USD-M futures. Direct analog of the landed BtcDailyVolumeSurgeBullishLong mechanism, applied to LINK. Already-landed LinkDailyMultiWeekTrendContinuationLong confirms LINK has clean trend-continuation behavior at daily timeframe. Single-instrument, OHLCV-only, single-condition entry, factory-approved instrument (LINKUSDT.BINANCE 1D).
Hypotheses
Direct analog of the landed BtcDailyVolumeSurgeBullishBarTrendFilterLong mechanism applied to LINK, whose clean daily trend-continuation behavior is confirmed by the landed LinkDailyMultiWeekTrendContinuationLong. Volume is the primary edge: an above-average volume day with a bullish close inside an established uptrend signals institutional-flow-driven continuation. The 100-bar SMA trend filter keeps the strategy out of downtrends and doubles as the primary exit, avoiding fee-churn chop. Sizing is capital-relative (30% of equity per position) with a $5 min-notional guard and LINK's 0.01 qty step. Kept on BINANCE USD-M futures (0.05% taker) rather than spot (0.10% taker) to halve fee drag while preserving a like-for-like analog to the proven BTC/ETH/SOL variants; leverage stays at 1.0 (no margin used).
Hypotheses
No edge on LINK — the BTC volume-surge mechanism fails to transfer (translation curse), and the failure is structural, not tunable. Realized track record is NEGATIVE: the +36.5% headline is a mark-to-market mirage (end_unrealized_pct 49.82 → realized ≈ -13.3%), and even that fake headline exists only because of 3 fat-tail outlier days (return_kurtosis 256, skew 14.5). Closed-trade reality: 12.5% win rate (4/32), 20 consecutive losses, expectancy -$416.79/trade, profit_factor 0.84, 52.3% max drawdown, negative annual returns in 4 of 7 years. A 12.5% win rate is a no-edge signal, not a parameter miscalibration — the entry (volume surge + bullish bar + uptrend) is simply wrong 7 out of 8 times on LINK, and no exit tuning or 2-hour optimization can invert a negative-expectancy entry; a sweep would only curve-fit the 3 lottery days into a Sharpe that collapses out-of-sample. Do not optimize. If the volume-surge continuation mechanism is worth pursuing, it belongs on the asset where it actually landed (BTC), not force-fit onto LINK.
Implementation
Long-only LINKUSDT daily strategy on BINANCE USD-M futures. Enters LONG when a daily volume surge (current volume >= vol_mult x average of the prior vol_lookback bars) coincides with a bullish bar (close > open) and an uptrend (close > SMA(trend_sma)). Exits on trend break (close < trend SMA) or an 8% catastrophic hard stop. OHLCV-only, single-instrument, leverage 1.0. calculate_signal returns relative volume (vol/avg - 1) continuously each bar; the surge+bullish+uptrend conjunction is thresholded in should_enter.
Backtest Review
Detector fires correctly: 32 clean long entries, no size-zero or min-notional drops; the volume-surge mechanism is implemented as described
Backtest Review
Trades match the hypothesis direction (all 32 long, volume-gated, trend-filtered)
Backtest Review
Headline +36.5% is illusory: end_unrealized_pct 49.82 means realized closed-trade return is ~-13.3% (an un-exited open position at backtest end)
Backtest Review
12.5% win rate (4/32) with 20 consecutive losses — the entry signal has negative edge on LINK
Backtest Review
Negative expectancy -$416.79/trade, profit_factor 0.84 (<1), CAGR -2.37%, Sharpe 0.024
Backtest Review
Catastrophic 52.3% max drawdown over a 2074-day underwater stretch
Backtest Review
Entire positive appearance depends on 3 fat-tail days (return_kurtosis 256, skew 14.5); removing them leaves a pure bleed
Backtest Review
The BTC volume-surge mechanism does not transfer to LINK — classic single-asset translation failure
Outcome Summary
This strategy translated the successful BTC daily volume-surge bullish-bar continuation mechanism onto LINKUSDT.BINANCE daily futures, entering long only when a volume surge, a bullish close, and a 100-day uptrend all aligned. The detector worked cleanly — 32 gated long entries with no size or notional drops — but the edge did not transfer: a 12.5% win rate, 20 consecutive losses, negative -$416.79/trade expectancy, profit factor 0.84, and a 52.3% drawdown, with the only positive-looking number being an unrealized mark-to-market headline that masked a realized loss of about -13.3%. The analyst abandoned it at the backtest-review gate, judging the failure structural rather than tunable and declining to spend optimization time curve-fitting three outlier days. It ended after a single iteration, never reaching optimization, final analysis, or risk review.
Outcome Summary
A mechanism that works on one asset does not automatically transfer to another — a 12.5% win rate with 20 consecutive losses signals no edge on LINK, not a parameter miscalibration, and optimizing it would only curve-fit the handful of lottery days.
Outcome Summary
The analyst issued an 'abandon' verdict at the pre-optimization backtest-review gate, ruling the negative edge structural rather than tunable, so the strategy never reached optimization, the post-optimization analyst decision, or risk review.
Outcome Summary
A long-only LINKUSDT daily futures strategy that bought on a volume surge (current volume ≥ 1.5× the prior 20-bar average) confirmed by a bullish bar and an uptrend (close > 100-day SMA), directly porting the landed BTC volume-surge continuation mechanism onto LINK.
Outcome Summary
Over 802 days of data it took 32 long trades with a 12.5% win rate (4/32), negative expectancy of -$416.79/trade, profit factor 0.84, Sharpe 0.024, CAGR -2.37%, and a 52.3% max drawdown; the +36.5% headline total return was a mark-to-market mirage (end_unrealized_pct 49.82 implies a realized return of roughly -13.3%) propped up by 3 fat-tail days (kurtosis 256, skew 14.5).
Backtest and paper results are hypothetical. Trading involves risk of loss.