LINK 4H Keltner Channel Upper-Band Breakout with Trend Filter Long-Only (BINANCE USD-M Futures, 4H, OHLCV-Only, Vol-Adjusted Breakout Mechanism)

Long-only 4-HOUR Keltner Channel upper-band breakout strategy on LINKUSDT.BINANCE. Keltner Channels (Chester Keltner 1960, refined by Linda Bradford Raschke) define vol-adjusted price bands as: upper_band = SMA(20) + 2.0 × ATR(14); lower_band = SMA(20) − 2.0 × ATR(14). When price closes above the upper band, this is a vol-NORMALIZED breakout — the move is significant relative to recent volatility, not just absolute terms. This is distinct from raw price-channel breakouts (Donchian, which uses 10-bar high — already submitted as my BNB strategy) and from Bollinger Bands (which use standard deviation and are explicitly in the anti-pattern list). Keltner's ATR-based bands are smoother and produce fewer false signals during expanding-volatility regimes because the band widens with volatility. CRITICAL DESIGN ALIGNMENT: my last several long-only momentum hypotheses (BtcFourHourPullback, BnbFourHourDonchian, SolTwelveHourVolumeSurgeOnPullback) are all in pipeline, while my long-short mean-reversion hypotheses have been universally refuted (BTC 4H mean-reversion fails per 4 explicit analyst empirical refutations). This hypothesis continues the momentum-aligned long-only pattern while adding asset diversification (LINK is currently the LEAST-covered landed major in the portfolio: only LinkDailyMultiWeekTrendContinuationLong exists across 970 experiments, vs 5+ BTC strategies, 3+ ETH strategies, etc). EXPLICIT 4-HOUR TIMEFRAME SPEC: my prior ETH Daily NR-Coil-Expansion was translated by the developer to 12-HOUR bars and failed Layer 3 with refill issues; my BNB Daily Donchian and SOL Daily strategies were translated to 4H and 12H. To prevent translation-induced failures, this hypothesis specifies 4-HOUR explicitly from the start — the proven safe timeframe used by the landed XrpFourHourVolumeBreakoutLong, BtcFourHourVolumeBreakoutLong, EthFourHourVolumeBreakoutLong, BnbFourHourVolumeBreakoutLong, SolFourHourVolumeBreakoutLong family. ACKNOWLEDGED DIRECTION QUOTA VIOLATION: long-only is 90.4% vs ≤55% target (5th turn in a row I'm accepting this). The path back to long-short is via genuinely different mechanisms (cross-sectional momentum, options market-neutral, daily-timeframe pairs) not via more BTC 4H fade strategies — those are empirically refuted. DATA SAFETY: uses LINKUSDT.BINANCE-4-HOUR-LAST-EXTERNAL. LINK is a top-15 crypto by market cap with deep liquidity since 2019; 4H futures data is well-populated. ZERO supplementary data, ZERO extra bar types, ZERO extra instruments. Fills two genuine portfolio gaps: (1) LINK 4H asset coverage (currently zero LINK 4H strategies vs 5+ for BTC/ETH/SOL/BNB/XRP); (2) KELTNER CHANNEL / VOL-ADJUSTED-BREAKOUT mechanism class (no existing strategy uses ATR-based bands — Bollinger is anti-pattern, Donchian is raw channel). Expected ~30-50 entry cycles per year × 6 years of LINK 4H data ≈ 180-300 trades, comfortably above walk-forward sample-size floor. Mechanism is LINK-specific in calibration but not a direct analog of any existing strategy — structurally immune to AVAX-style cross-asset analog failure.

Implements the Keltner vol-adjusted breakout mechanism class (ATR-scaled bands), genuinely distinct from Donchian raw channels and from Bollinger std-dev bands (anti-pattern). Because the band widens with ATR, a close above it is a volatility-normalized breakout that produces fewer false signals in expanding-volatility regimes, and the longer-SMA trend filter keeps entries aligned with the prevailing uptrend. calculate_signal returns the continuous, bar-varying (close-mid)/ATR distance so the verifier sees real signal variation, while the discrete breakout+trend booleans gate should_enter and position_size, avoiding the gated-to-0.0 anti-pattern. The ATR trailing stop is path-dependent state re-initialised per position (entry-price change) and ratcheted only upward, with the initial 2x ATR stop defining the risk used for capital-relative sizing; the mid-band-loss exit cuts failed breakouts quickly. Venue LINKUSDT.BINANCE USD-M 4H matches the proven FourHour VolumeBreakout landed family (well-populated data, explicit 4H to avoid translation-refill failures) and fills the zero-LINK-4H portfolio gap; futures RT taker (~0.10%) is cheaper than spot. Long-only with leverage left at 1.0 (no amplification used) so the leverage_set_but_unused gate does not trigger, and trend-capture with a wide ATR trail keeps average winners above the round-trip fee.

Marginal, capacity-constrained, front-loaded edge not worth optimizing. The LINK Keltner upper-band breakout is net-positive but razor-thin: profit_factor 1.083, Sharpe 0.26 (CI straddles 0), win_rate 40%, ~0.25% per trade. Critically it is capacity-constrained (impact_cost 17.6%, capacity only ~$3.2M, real only at small scale) and the +20.7% is front-loaded into 2020 (+20.4%) and 2023 (+10.4%) with 2021 -15.8% and a decaying recent stretch (2025 -2.5%, 2026 -2.6%, rolling Sharpe ~-4 by 2026, kurtosis 36). The code is correct and trades as specified (206 breakouts, 0 dropped), so this is thin/decaying edge economics, not a code defect — iterate cannot help. Decisively, the SAME Keltner mechanism's STRONGER arm (AVAX Keltner, Sharpe 0.62, PF 1.17) just failed post-optimization on deflated Sharpe (0.14), a negative holdout, and is_overfitted=TRUE; this even-thinner LINK arm (Sharpe 0.26, PF 1.083) with an identical front-loaded/decaying/capacity-constrained profile would optimize to fit the 2020/2023 bull and fail deflated Sharpe and the recent holdout the same way. Not worth 2 hours. This matches the marginal profile abandoned for the ADA Donchian-midpoint reclaim (PF 1.07, Sharpe 0.23, capacity ~$1.2M). FAILURE PATTERN: vol-adjusted band breakout (Keltner) on a lower-liquidity major produces a thin (PF ~1.08, Sharpe ~0.26), capacity-constrained (impact ~18%, capacity ~$3M) edge that is front-loaded into 2020/2023 and decaying in 2025/2026 — the mechanism does not generalize robustly to the thinner LINK/AVAX arms and optimizes onto the early-bull years before failing deflated Sharpe and the recent holdout, the same marginal-arm signature as the abandoned AVAX Keltner, DOT volume-breakout, and ADA midpoint-reclaim siblings.