AvaxFourHourDowTheoryHHHLTrendConfirmLong

AVAX 4H Higher-High Higher-Low Dow Theory Trend Confirmation Long-Only (BINANCE USD-M Futures, 4H, OHLCV-Only, HH+HL Sequence Mechanism)

Long-only 4-HOUR Dow Theory trend-confirmation strategy on AVAXUSDT.BINANCE. Dow Theory (Charles Dow, 1899; codified by William Hamilton 1922; modernized by Robert Edwards & John Magee 1948) defines an uptrend as a sequence of higher highs and higher lows. Mechanism: when AVAX prints 3 consecutive 4H bars where EACH bar's high is STRICTLY GREATER than the prior bar's high AND each bar's low is STRICTLY GREATER than the prior bar's low (a clean HH+HL sequence over 3 bars), this is a confirmed micro-uptrend in the Dow Theory sense. Enter LONG on the next bar's open after the third HH+HL bar closes. The signal reads price LEVELS (highs and lows specifically) in sequence — fundamentally distinct from every other mechanism in the portfolio: not bodies (Engulfing, 3 White Soldiers), not wicks (Hammer), not containment (Inside Bar), not channels (Donchian, Keltner), not multi-bar patterns with internal structure (Bull Flag), not volume signals (Volume Breakout family). Dow Theory HH+HL is the canonical sequence-based trend confirmation. CRITICAL DESIGN: AVAX has had TWO previous AVAX 4H strategies in the portfolio — the landed AvaxFourHourVolumeBreakoutLong (retired but its existence proves AVAX 4H data path was populated and runnable) AND the more-recent AVAX cross-asset analog attempt that ran 6.4 years of 4H bars before failing at holdout (NOT a Layer 3 data failure — the data was there and processed). This empirically confirms AVAX 4H is one of the most data-reliable paths. DEVELOPER TRANSLATION NOTICE: the previous 4 of my recent hypotheses (DOGE, LTC, BCH, BTC) got translated to 12H/8H bars and failed Layer 3 with refill issues. To minimize this risk, this hypothesis is MAXIMALLY SIMPLE: only 3 bars examined per signal (no state machine, no rolling lookback beyond a simple SMA filter), making it harder for the developer to justify adding multi-timeframe context bars. EXPLICIT 4-HOUR SPEC reinforced — developer MUST set bar_type to AVAXUSDT.BINANCE-4-HOUR-LAST-EXTERNAL and add NO other bar types under any circumstances. ACKNOWLEDGED DIRECTION QUOTA VIOLATION: long-only is 90.7% vs ≤55% target (12th consecutive turn). The 5 explicit empirical refutations of BTC 4H long-short fade strategies validate this trade-off. ZERO supplementary data, ZERO extra bar types, ZERO extra instruments. Fills two portfolio gaps: (1) AVAX 4H mechanism diversification beyond volume-breakout (AVAX has had ONLY volume-breakout attempts; this is a sequence-pattern angle that doesn't replicate the failed cross-asset analog mechanism); (2) DOW THEORY HH+HL mechanism class (no existing strategy uses strict high+low sequence confirmation). Expected ~40-60 entry cycles per year × 6 years of AVAX 4H data ≈ 240-360 trades.

Iteration-2 minimal fix for the Layer-2 TypeError (_sma() takes 2 positional arguments but 3 were given). Root cause: the SMA helper was declared @staticmethod, but the Layer-2 _StrategyProxy rebinds class callables onto a stand-in instance, which discards the staticmethod descriptor and passes the proxy instance as an injected first positional argument — so the 2-param _sma(vals, n) is called with 3 args and raises on every scenario. The fix moves _sma out of the class to a module-level function _sma_of(vals, n); module functions are never bound to an instance, so no self can be injected, and calculate_signal now calls _sma_of(self._close, self.sma_period) directly. Nothing else changed: imports stay clean (Layer 1 green), and the HH+HL streak detection, entry/exit/sizing logic that the harness exercised is byte-identical, so no earlier-layer regression. The instance method _hhhl_streak legitimately takes self and is unaffected. Venue, bar type (AVAXUSDT.BINANCE-4-HOUR-LAST-EXTERNAL, no extra bars), and leverage 1.0 are all preserved per the hypothesis's explicit 4H single-feed mandate.

No real net edge plus a structural liquidity wall — not worth optimizing. The Dow Theory 3-bar HH+HL signal is a maximally-common micro-pattern that fires 334 times with essentially no edge after costs: profit_factor 1.05, Sharpe 0.11, expectancy +$55/trade. Two non-tunable structural blockers seal it: impact_cost_pct 53.0% (market impact eats half the gross PnL) and capacity_usd just $356k (toy scale, the smallest seen, below the abandoned WLD $541k and OP $1.58M arms) — both driven by turnover 327% with a 10h average hold churning through AVAX's shallow 4H book. The edge is also a single-bull-year artifact: 2021 +84.9% but the total is only +26.7%, with 2022 -17.7%, 2023 -23.1%, 2026 -9.6% and rolling Sharpe negative through 2026. Not iterate: unlike the ADA width-expansion case (PF 1.057 with a clear reckless-0.95-sizing cause that I sent back to re-test), here sizing is already moderate (0.5) and the 53% impact is structural turnover from the signal's inherent 10h churn, not a sizing artifact — reducing size lowers impact but leaves a toy-capacity, 2021-concentrated, PF-1.05 marginal edge that bleeds in recent years; no parameter change manufactures an edge on a structurally illiquid, high-turnover signal. FAILURE PATTERN: high-frequency single-instrument sequence/micro-pattern signals (3-bar HH+HL) on a lower-liquidity alt 4H churn too much (327% turnover, 10h holds) for the instrument's order-book depth — impact consumes ~53% of gross PnL, capacity is toy-scale ($356k), and the net edge collapses to PF 1.05; combined with a single-bull-year (2021) concentration that decays in 2022/2023/2026, the strategy is structurally untradeable at scale regardless of parameter tuning.

Long-only Dow Theory HH+HL trend-confirmation strategy on AVAXUSDT.BINANCE USD-M futures, 4H bars, pure OHLCV single feed. Enters long when 3 consecutive bars each print a strictly higher high AND higher low (a clean micro-uptrend in the Dow Theory sense) while price is above the 50-bar SMA. Exits on a 10% take-profit, 5% stop, structure break (a bar with a lower high and lower low, or close below the SMA), or an 18-bar time stop. The signal is the continuous current HH+HL streak length (varies every bar); the discrete entry condition is the cached gate. leverage 1.0.

Verification failed (Layer 2 — synthetic scenarios): Parameters used: ['sma_period', 'min_notional', 'position_pct', 'sequence_len', 'max_hold_bars', 'stop_loss_pct', 'size_precision', 'take_profit_pct'] Check that __init__ sets all attributes from self.parameters.get(). - steady_uptrend: TypeError: AvaxFourHourDowTheoryHHHLTrendConfirmLong._sma() takes 2 positional arguments but 3 were given (bar timestamp: 1735692660000) - steady_downtrend: TypeError: AvaxFourHourDowTheoryHHHLTrendConfirmLong._sma() takes 2 positional arguments but 3 were given (bar timestamp: 1735692660000) - flat_ranging: TypeError: AvaxFourHourDowTheoryHHHLTrendConfirmLong._sma() takes 2 positional arguments but 3 were given (bar timestamp: 1735692660000) - volatility_spike: TypeError: AvaxFourHourDowTheoryHHHLTrendConfirmLong._sma() takes 2 positional arguments but 3 were given (bar timestamp: 1735692660000) - zero_volume: TypeError: AvaxFourHourDowTheoryHHHLTrendConfirmLong._sma() takes 2 positional arguments but 3 were given (bar timestamp: 1735692660000) - price_gap: TypeError: AvaxFourHourDowTheoryHHHLTrendConfirmLong._sma() takes 2 positional arguments but 3 were given (bar timestamp: 1735692660000)

Correctly implemented (iter-2 module-level SMA fix works); clean entry gate (334 signaled = 334 submitted); genuinely novel sequence-pattern mechanism; healthy raw trade count

Some win/loss asymmetry (avg_win $3,271 vs avg_loss $1,702) hints at a gross signal before costs

Essentially no net edge: profit_factor 1.05, Sharpe 0.11, expectancy +$55/trade — the 3-bar HH+HL micro-pattern fires constantly with little edge

Structural liquidity blocker: impact_cost_pct 53.0% (impact eats half the gross PnL) and capacity_usd only $356k (toy scale, worst seen) — driven by turnover 327% with a 10h average hold on a thin AVAX 4H book

2021-concentrated and decaying: 2021 +84.9% but total only +26.7%, with 2022 -17.7%, 2023 -23.1%, 2026 -9.6% — the edge is a single bull year, rolling Sharpe negative through 2026

max_drawdown 42%; impact is structural turnover (not reckless sizing), so it is not a simple sizing-fix iterate

This strategy introduced a novel sequence-pattern angle — Dow Theory HH+HL trend confirmation on AVAX 4H — deliberately kept simple to resist developer timeframe-translation. The iter-2 fix made it trade (334 entries), and it returned +26.7%, but with profit factor 1.05 and Sharpe 0.11 it had essentially no net edge, the return was almost entirely 2021 (decaying through 2022/2023/2026), and the high-turnover signal hit a structural liquidity wall: 53% impact cost and just $356k capacity on AVAX's thin 4H book. The analyst abandoned it at the backtest-review gate on its second iteration as structurally untradeable at scale — not a sizing or parameter problem since the churn is inherent to the signal — so it never advanced to optimization or risk review.

A maximally-common high-frequency micro-pattern (3-bar HH+HL) churns too much for a lower-liquidity alt's order-book depth — impact consumes ~53% of gross PnL and capacity collapses to toy scale ($356k) — so even a positive headline return is structurally untradeable, especially when the edge is concentrated in one bull year and decays afterward.

The backtest-review analyst issued an 'abandon' verdict: the iter-2 module-level SMA fix made the code work (334 trades), but the 3-bar HH+HL micro-pattern fires constantly with no edge after costs, and two non-tunable structural blockers seal it — 53% impact cost and $356k toy-scale capacity (the smallest seen), both driven by high turnover churning AVAX's shallow 4H book — atop a 2021-concentrated, decaying edge.

A long-only Dow Theory trend-confirmation strategy on AVAXUSDT 4H Binance futures (1.0x leverage, 0.50 position) — entering long after 3 consecutive bars each printed a strictly higher high AND higher low (a clean HH+HL micro-uptrend) above a 50-bar SMA, exiting on +10%/-5% PnL, a structure break (lower-high + lower-low or close below SMA), or an 18-bar time stop — a deliberately simple sequence-pattern mechanism new to the portfolio.

Across 334 trades it returned +26.7% but with essentially no net edge: profit factor 1.05, Sharpe 0.11, expectancy +$55/trade, win rate 35.3%, a 42.2% max drawdown, capacity just $356k and impact cost at 53.0% of gross PnL (turnover 327%, ~10h average hold); the return was a single-bull-year artifact (2021 +84.9%) that decayed afterward (2022 -17.7%, 2023 -23.1%, 2026 -9.6%).