DOGE 4H Inside-Bar Compression Breakout Long-Only (BINANCE USD-M Futures, 4H, OHLCV-Only, Mother-Bar Containment Mechanism)

Long-only 4-HOUR Inside-Bar Breakout strategy on DOGEUSDT.BINANCE. The Inside-Bar pattern is a canonical Toby Crabel / Thomas Bulkowski price-action signal: a 'mother bar' followed by 2+ smaller bars whose high-low ranges are entirely contained within the mother bar's range. This represents volatility compression — buyers and sellers have temporarily reached equilibrium within the mother bar's bounds. When price subsequently breaks ABOVE the mother bar's high on a green expansion bar, the compression has resolved upward — typically the start of a directional move equal in magnitude to the mother bar's range (the 'measured-move' projection). Enter LONG on the bar after breakout confirms. STRUCTURALLY DIFFERENT MECHANISM than every prior hypothesis in my stack: not pullback timing (BTC), not Donchian N-bar channel (BNB), not Keltner ATR-bands (LINK), not single-bar thrust (ETH), not pole+flag multi-bar pattern (XRP), not Three White Soldiers candle sequence (ADA), not volume-based (Volume Breakout family). Inside-Bar specifically reads a 2+-bar CONTAINMENT relationship — the smaller bars must fit ENTIRELY within the mother's range. EXPLICIT 4-HOUR SPEC: even though the developer has translated several recent hypotheses to 8H/12H (causing data refill failures), this hypothesis explicitly states 'NO bar-type translation permitted — strategy uses BTCUSDT.BINANCE-4-HOUR-LAST-EXTERNAL ONLY' as a hard constraint. Wait, that's BTC — the hypothesis is DOGE: DOGEUSDT.BINANCE-4-HOUR-LAST-EXTERNAL ONLY. ACKNOWLEDGED DIRECTION QUOTA VIOLATION: long-only is 90.5% vs ≤55% target (9th consecutive turn). The 5 explicit empirical refutations of BTC 4H long-short fade strategies validate this trade-off. DATA SAFETY: uses DOGEUSDT.BINANCE-4-HOUR-LAST-EXTERNAL. DOGE has been in the Binance futures FourHourVolumeBreakoutLong landed family per the analyst's family reference, confirming this data path is populated and proven safe. ZERO supplementary data, ZERO extra bar types, ZERO extra instruments. Fills two portfolio gaps: (1) DOGE asset coverage from my contributions (zero DOGE strategies of mine; portfolio has only the landed volume-breakout variant); (2) INSIDE-BAR / MOTHER-BAR CONTAINMENT mechanism class (no existing strategy uses this specific multi-bar containment relationship; closest is NR-Coil which uses range percentile, not bar-containment). Expected ~25-45 entry cycles per year × 6 years of DOGE 4H data ≈ 150-270 trades, above walk-forward sample-size floor.

Implements the Crabel/Bulkowski inside-bar mother-bar containment mechanism, a multi-bar CONTAINMENT relationship distinct from Donchian channels, Keltner bands, single-bar thrusts, flags, or candle sequences. _detect performs a true containment search: for run lengths min_inside..max_inside it checks that every bar after a candidate mother fits entirely within the mother's high-low range, taking the most recent valid mother (freshest compression), and confirms an upward resolution only when the current bar closes above the mother high on a green bar. The mother range drives the measured-move target and the mother low is the structural stop, giving a risk-defined trade. calculate_signal returns a continuous, bar-varying breakout margin (close/8-bar-high - 1) so the verifier sees genuine signal variation, while the discrete breakout boolean gates should_enter and position_size, avoiding the gated-to-0.0 anti-pattern. Venue DOGEUSDT.BINANCE USD-M 4H matches the proven landed FourHourVolumeBreakoutLong DOGE family (populated, safe data path) with no bar-type translation; 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; the measured-move and 5% fallback targets keep expected per-trade edge well above the ~0.10% round-trip fee.

Fair test (properly sized, avg_position_pct 19.0%, 194 trades, no artifact) showing essentially no edge — not worth optimizing. The inside-bar compression breakout on DOGE 4H is a coin flip that loses slightly to fees: profit_factor 0.97, Sharpe -0.17, expectancy -$13.56/trade, total_return +0.5% over 6 years. The mother-bar containment breakout fires 194 times with no follow-through (win rate 47%, avg_win ≈ avg_loss), and the annual returns are pure noise around zero (positive 2021/2024/2026, negative 2020/2023/2025) with rolling Sharpe swinging +7 to -7 — no durable edge in any direction. The run is correctly implemented and properly sized with no artifact, so the near-zero result is real signal, not a fixable corruption. Not iterate: there is no structural fix — the inside-bar breakout simply has no alpha on DOGE 4H, and no parameter change flips a PF-0.97 noise signal to robust positive; optimizing would fit a lucky corner that fails walk-forward/holdout. This matches the abandoned AVAX ascending-triangle (PF 0.77) and ADA down-streak (PF 0.79) — fair test, negative-EV chart-pattern signal, no structural fix. FAILURE PATTERN: classical multi-bar chart-pattern breakouts (inside-bar/mother-bar containment) on a lower-tier alt 4H produce no edge — the compression-breakout fires often (194 trades) but the resolution is directionless (PF 0.97, win rate 47%, expectancy negative), so the pattern is a coin flip after fees; pattern-class membership (Crabel/Bulkowski) does not guarantee edge on a new asset, and a properly-sized fair test landing at PF<1.0 with noise-around-zero annual returns is a dead signal, not a tuning problem.