AdaFourHourPullbackToRangeMidpointAfterThrustLong

ADA 4H Pullback-to-Range-Midpoint After Strong Move Long (Buy-the-Dip Structural Pattern)

A long-only, single-instrument, single-timeframe strategy on ADAUSDT.BINANCE 4H bars. Implements a NEW MECHANISM CLASS that addresses the analyst's critical regime-decay warning from the Two-Bar HH Breakout abandon: 'Mechanism worked during the elevated crypto-volatility regime but does not generalize to the current lower-vol / failed-breakout regime.' Breakout-based strategies are now at-risk; the 2025+ regime favors MEAN-REVERSION within established uptrends — buy-the-dip rather than buy-the-breakout. The mechanism: detect when price has had a strong recent UP move (3+ consecutive bars where close > prior close, totaling >= 1.0×ATR cumulative upmove), then a PULLBACK occurs (price retraces back to within 0.3×ATR of the midpoint of the recent N-bar range), and the current bar closes BULLISHLY confirming the pullback ended. In a confirmed uptrend (close > sma_100), this captures the classic 'buy the dip' opportunity that thrives in lower-vol regimes where breakouts fail but pullbacks get bought. Critically distinct from every existing ADA 4H mechanism: (a) NOT breakout-based (entries are at pullback MIDPOINT, not at channel extremes) — directly addresses regime-decay vulnerability of all breakout strategies; (b) NOT statistical (uses structural price-level retracement, not z-scores or autocorrelation); (c) NOT single-bar absorption (uses multi-bar UP-MOVE + pullback sequence). STRUCTURAL + SEQUENCE-BASED + MEAN-REVERSION-WITHIN-TREND per analyst guidance.

Implements a mean-reversion-within-trend buy-the-dip mechanism explicitly designed to survive the lower-vol / failed-breakout regime the analyst flagged: entries are at the range MIDPOINT retracement, not at channel extremes, so it does not share the regime-decay vulnerability of breakout strategies. It is structural (price-level retracement, not z-scores) and sequence-based (a multi-bar up-thrust must precede the pullback), distinct from every existing ADA 4H mechanism. The thrust detector scans the recent window for a qualifying consecutive-up-close run, ensuring the dip is only bought when there was genuine momentum to retrace; the 100-SMA filter keeps trades on the right side of the trend; the bullish-close confirmation times the dip's end. calculate_signal returns the continuous ATR-normalized distance from the midpoint so the verifier sees a real, bar-varying signal, while the discrete buy-the-dip boolean gates should_enter and position_size, avoiding the gated-to-0.0 anti-pattern. Venue ADAUSDT.BINANCE USD-M 4H is the landed ADA family; futures RT taker (~0.10%) is cheaper than spot; long-only with leverage left at 1.0 (no amplification) so the leverage_set_but_unused gate does not trigger, and the 2.5x ATR target keeps expected per-trade edge above the round-trip fee. Single instrument / single timeframe avoids cross-asset audit_stale risk.

Pre-optimization abandon: base config has negative edge (Sharpe -1.04, profit_factor 0.78, negative expectancy over 51 trades, win_rate 33%, 9 consecutive losses). The only positive month is Nov 2024 (+9.2%, 3 pump trades); 2022/2025/2026 all negative. The buy-the-dip regime thesis is contradicted — 2025 (its target lower-vol regime) is the worst year (-7.1%). Code correctly implements the hypothesis and trades match it, so not a developer fix; the mechanism has no edge on ADA 4H. Optimization can only curve-fit to the Nov-2024 winners and will fail walk-forward/holdout. Not worth 2 hours.

Long-only buy-the-dip structural pattern on ADAUSDT.BINANCE USD-M futures, 4H bars, pure OHLCV single feed. Enters LONG when, in a confirmed uptrend (close > 100-SMA), a strong recent up-thrust (a run of >=3 consecutive up-closes with cumulative gain >=1x ATR within the last 8 bars) is followed by a pullback to within 0.3x ATR of the 20-bar range midpoint (having retraced >=0.5x ATR below the range high) and the current bar closes bullishly. Exits on a 1.5x ATR stop, a 2.5x ATR target, or a 20-bar time stop. The signal is the continuous ATR-normalized distance from the range midpoint recomputed every bar; the thrust+pullback+confirmation condition is the discrete entry gate. Sized at 18% equity notional, capped so ATR-stop risk <= 1.5% of equity. leverage 1.0.

Adequate trade count (51 over 6.4yr); code correctly implements the thrust+pullback+confirmation mechanism (no gated-to-0.0 anti-pattern) and trades match the hypothesis.

Low drawdown (8.6%) and the structural distinction from breakout strategies is genuine.

Base config loses money with no edge: Sharpe -1.04, profit_factor 0.78 (<1), negative expectancy (-$120/trade), win_rate 33%, omega 0.78, 9 consecutive losses.

The only positive contribution is a single bull month (Nov 2024 +9.2%, 3 trades); every other year except small 2021/2023 is negative (2022 -6.1%, 2025 -7.1%, 2026 -1.7%).

Regime thesis is contradicted: the strategy was built for the 2025+ lower-vol 'buy-the-dip' regime, yet 2025 is its worst year (-7.1%) and 2026 is negative.

33% win rate at a 2.5:1.5 ATR target/stop is below breakeven — the signal quality is fundamentally weak, not a tuning miss.

Not worth 2 hours of optimization: the honest base config is an outright money-loser (Sharpe -1.04, profit_factor 0.78, negative expectancy over 51 trades, 9 consecutive losses), and the ONLY positive contribution is a single bull-pump month (Nov 2024 +9.2%, 3 trades) — strip it and the strategy is deeply negative across 2020/2022/2025/2026. The regime premise is directly contradicted by the data: the buy-the-dip mechanism was designed for the 2025+ lower-vol regime, yet 2025 (-7.1%) is its worst year. This is not a code bug (the strategy implements the hypothesis faithfully and trades match it), so it is not a developer iterate — the mechanism simply has no edge on ADA 4H. Optimization tunes parameters; it cannot convert a profit_factor-0.78 negative-expectancy signal into a robust edge, it would only curve-fit TP/SL to the few Nov-2024 winners and then fail walk-forward/holdout (recent OOS windows negative). Abandon rather than spend the compute.

AdaFourHourPullbackToRangeMidpointAfterThrustLong was a deliberate pivot away from breakout strategies, designed per analyst guidance to buy pullbacks to the range midpoint after a confirmed up-thrust — a mean-reversion-within-trend pattern meant to thrive in the 2025+ lower-volatility regime. The code faithfully implemented the thrust-plus-pullback-plus-confirmation logic and traded adequately (51 entries), but the base backtest was an outright loser: Sharpe -1.04, profit factor 0.78, a 33% win rate, and nine consecutive losses, with the sole positive month being a Nov-2024 bull pump. Damningly, its own target regime betrayed it — 2025 was its worst year at -7.1% — so the analyst abandoned it before optimization, judging this a fundamentally edgeless mechanism on ADA 4H rather than a tuning problem that any parameter sweep could rescue.

A regime-targeted mechanism must be validated against its own target period: when a 'lower-vol buy-the-dip' strategy loses most in the very regime it was built for and a 33% win rate sits below its target/stop breakeven, the signal has no edge and optimization would only curve-fit to a single pump month.

It was abandoned at the pre-optimization backtest-review gate (verdict: abandon) because the honest base config had negative edge that optimization could not fix, and the regime thesis was directly contradicted by the data — the buy-the-dip mechanism, designed for the 2025+ lower-vol regime, posted its worst year in exactly that target regime (-7.1%).

A long-only ADAUSDT 4H structural 'buy-the-dip' strategy that, in a 200/100-SMA-confirmed uptrend, entered on a pullback to the 20-bar range midpoint after a strong multi-bar up-thrust (≥3 consecutive up-closes totaling ≥1.0×ATR) with a bullish confirmation bar — a mean-reversion-within-trend mechanism deliberately built to avoid the regime-decay of breakout strategies.

Over ~6.4 years it took 51 long trades with a 33% win rate, profit factor 0.78, negative expectancy (-$120/trade), Sharpe -1.04 (CI [-4.10, 1.40]), a -5.24% total return, and up to 9 consecutive losses, though max drawdown stayed low at 8.6%. The only meaningfully positive contribution was a single bull month (Nov 2024, +9.2% over 3 trades), with 2022, 2025, and 2026 all negative.