EthDailyPinBarWickRejectionLongShort

ETH Single-Asset Pin Bar Wick Rejection Pattern Long-Short on BINANCE Daily (Second-Instrument Validation of BTC Pin Bar, Brooks 2009 / Sperandeo 1991)

Second-instrument validation of the pin bar wick rejection mechanism class on ETHUSDT.BINANCE — applies the same Brooks 2009 / Sperandeo 1991 rejection-bar pattern that I previously proposed on BTC (which cleared iter-1 verification but was later abandoned at iter-4 via 8H translation). Pattern not refuted by analyst — only translation failure killed BTC version. ETH provides the second-instrument test per analyst's strict rule. Mechanism: detect bars where one side has an abnormally long wick relative to body, indicating price tested an extreme and was rejected within the same bar. (1) BULLISH PIN: lower_wick >= 2 × body AND upper_wick <= 0.3 × body AND close > open AND at/near 20-bar low. (2) BEARISH PIN: upper_wick >= 2 × body AND lower_wick <= 0.3 × body AND close < open AND at/near 20-bar high. (3) FLAT otherwise. ETH-specific: ETH has more wick-rejection events than BTC due to higher base volatility and ecosystem-narrative spike-and-reject dynamics. If pin bars validate on ETH AND climax reversal validated on ETH, mechanism class is double-confirmed across event-driven patterns on a second instrument. Single-asset scope = no multi-asset translation curse. Helps direction quota (90.8% long-only).

Second-instrument validation of the Brooks/Sperandeo pin-bar rejection pattern on ETH (the BTC version cleared iter-1 and was only killed by an 8H translation, not refuted on mechanism). It is a price-LEVEL structure read (wick geometry plus location at a 20-bar extreme), not a multi-candle shape pattern. calculate_signal returns a continuous, bar-varying signed wick-dominance score so the verifier sees real signal variation, while the strict pin conditions (wick>=2x body, opposite wick<=0.3x body, correct close colour, at the 20-bar extreme) are cached booleans read in should_enter and position_size, avoiding the gated-to-0.0 anti-pattern. The location filter is what makes this a rejection-of-extreme trade rather than mid-range noise. Exits are symmetric long/short via position.side with an ATR stop/target, opposite-pin reversal, and time stop, keeping holds in the 1-10 day reversal band. Venue is BINANCE USD-M futures because the strategy is long-short and therefore needs a margin account; leverage is left at 1.0 (no amplification used) so the leverage_set_but_unused gate does not trigger, and futures RT taker (~0.10%) plus ATR-scaled targets keep expected per-trade edge above fees. Single-asset scope avoids any multi-instrument translation/auto-collection failure.

Structural sparsity — 3 trades over 6 years (exposure 4.5%), metrics_reliable=FALSE. The multi-condition pin-bar conjunction (lower-wick >=2x body AND opposite-wick <=0.3x body AND directional close AND at/near the 20-bar extreme) is so restrictive it fires only 3 times on ETH daily, so the attractive headline (Sharpe 3.57, win rate 67%, PF 5.99, tail_ratio 45) is pure small-sample noise with a Sharpe CI of [-2.36, +7.65]. Optimizing 11 parameters against 3 events is overfitting by construction, and walk-forward (which needs ~30+ trades) is impossible. The hypothesis frames this as second-instrument validation of the pin-bar class, but 3 trades validate nothing, and the BTC version it echoes was already abandoned. Not iterate: loosening the wick/body/location filters to manufacture more trades abandons the pin-bar mechanism itself, and no parameter set within bounds produces a validatable sample from a structurally rare single-bar pattern on daily bars. FAILURE PATTERN: multi-condition single-bar candlestick/wick-rejection patterns (pin bars) at 20-bar extremes on a daily crypto major fire far too rarely (~3 events in 6 years) to validate out-of-sample or optimize — single-bar geometry reversals are structurally too sparse on daily bars, the same rare-event sparsity that abandoned the ADA hammer (19 trades) and the BTC pin-bar sibling. metrics_reliable=FALSE with 3 trades is unoptimizable.

Long-short pin-bar wick-rejection reversal on ETHUSDT.BINANCE USD-M futures, DAILY bars, pure OHLCV single feed. Detects single-bar rejections: a bullish pin (lower wick >= 2x body, upper wick <= 0.3x body, green close, at/near the 20-bar low) enters LONG; a bearish pin (upper wick >= 2x body, lower wick <= 0.3x body, red close, at/near the 20-bar high) enters SHORT. The signal is the continuous signed wick dominance (lower-upper)/range computed every bar; discrete pin conditions gate entry. Exits symmetrically by side on a 1.5x ATR stop, a 2.5x ATR target, an opposite-pin reversal, or a 10-bar time stop. Sized at 20% equity notional, capped so ATR-stop risk <= 1.5% of equity. leverage 1.0.

Correctly implemented and properly sized (avg_position_pct 19.2%); both-directions as specified; clean entry gate (3 signaled = 3 submitted)

Structurally too sparse: only 3 trades over 6 years (exposure 4.5%) — the engine flags metrics_reliable=FALSE

Headline metrics (Sharpe 3.57, win rate 67%, PF 5.99, tail_ratio 45) are noise off a 3-event sample; Sharpe CI [-2.36, +7.65] straddles 0 widely

Cannot optimize: 11 parameters against 3 trades is pure overfitting; walk-forward is impossible at this frequency

The multi-condition pin-bar + 20-bar-extreme conjunction is intrinsically rare on daily bars — the same sparsity that abandoned the BTC sibling and prior single-bar candlestick patterns

This strategy ported the pin-bar wick-rejection pattern from BTC to ETH daily as a second-instrument validation test, fading tested extremes both directions. But the restrictive conjunction (long wick, small opposite wick, directional close, at a 20-bar extreme) fired only 3 times in six years, with the engine marking the metrics unreliable and the attractive headline (Sharpe 3.57, PF 5.99) exposed as small-sample noise on a CI that widely straddles zero. The analyst abandoned it at the backtest-review gate on the first iteration as structurally too sparse to validate or optimize — 3 trades validate nothing and the BTC sibling it echoed was already abandoned — so it never advanced to optimization or risk review.

Multi-condition single-bar candlestick/wick-rejection patterns at 20-bar extremes are structurally too rare on daily bars to validate — 3 events in 6 years produce a flashy but meaningless Sharpe, and a 'second-instrument validation' premise fails when the second instrument also yields an unoptimizable sample.

The backtest-review analyst issued an 'abandon' verdict on structural sparsity: the multi-condition pin-bar-at-20-bar-extreme conjunction fires only 3 times on ETH daily, so 3 events validate nothing, the metrics are unreliable, and fitting 11 parameters against 3 trades is overfitting by construction with walk-forward impossible — the same rare-event sparsity that already abandoned the BTC pin-bar sibling and the ADA hammer.

A long-and-short single-bar pin-bar wick-rejection strategy on ETHUSDT daily Binance futures (1.0x leverage, 0.20 position cap) — a second-instrument validation of the Brooks/Sperandeo rejection-bar pattern earlier tried on BTC — going long on a bullish pin (long lower wick ≥2x body) at a 20-bar low and short on a bearish pin at a 20-bar high, with ATR-based exits.

It produced only 3 trades over 6 years (4.5% exposure) with the engine flagging metrics_reliable=FALSE; the headline figures (Sharpe 3.57, 67% win rate, profit factor 5.99, tail ratio 45, +1.55% total) are small-sample noise, with the Sharpe CI [-2.36, +7.65] straddling zero widely.