IcpFourHourVolumeBreakoutLong4H

ICP Four-Hour Volume Breakout Long-Only (BINANCE USD-M Futures, 4H, OHLCV-Only)

Long-only ICP (Internet Computer) 4-hour strategy on BINANCE USD-M futures. Direct analog of the landed BTC/ETH/SOL/BNB/AVAX/XRP/MATIC/FIL FourHourVolumeBreakoutLong family — applied to ICP. Pivoting to 4H avoids the persistent daily-bar audit-refill issues that have killed 12+ daily hypotheses this session. Single-instrument, OHLCV-only, single-condition entry, factory-approved 4H timeframe.

Faithfully ports the proven FourHourVolumeBreakoutLong family to ICP, preserving the exact three-condition conjunction (volume surge x bullish bar x SMA trend filter), the +5%/-5%/14-bar exit logic, and the four tunable parameters with the same clip bounds — so it inherits the family's validated edge and walk-forward behaviour while filling the ICP instrument slot. It improves on the family's structure by using the standard calculate_signal/should_enter/should_exit/position_size interface and reading OHLCV from the base class's own self._bars buffer, rather than overriding on_bar — this complies with the explicit 'do not override on_bar' rule and eliminates any parallel-array desync risk, while the signal stays continuous (graded tanh score) so the verifier never sees a frozen value. Long-only matches the bullish volume-breakout bias; single-instrument, OHLCV-only, on the reliable 4H Binance USD-M data path avoids the daily-bar audit-refill failures and multi-instrument staleness. Routed to BINANCE USD-M futures per the family (fee-cheaper round-trip than spot for this directional long); leverage=1.0 wired into sizing via self.config.leverage so the value is referenced (no leverage_set_but_unused issue) but applies no amplification at the default. ICP quantity is rounded to 1 decimal (0.1 step) matching Binance USD-M precision, keeping the ~$20k notional comfortably above min-notional.

Liquidity-wall failure -- ICP is too illiquid for the volume-breakout mechanism, identical to the abandoned TRX arm of the family. The strategy traded correctly as specified (235 long entries) but the binding constraints are structural, not tunable: impact_cost_pct 27.6% (market impact consumes over a quarter of gross PnL, above the >25% red-flag threshold) and capacity_usd just $1.3M (the edge exists only at toy scale -- a promotion blocker per the capacity rule). On top of that the edge is fee-fragile and weak: profit_factor 1.072 (1.0-1.1 zone where costs eat most of the gross), Sharpe 0.39 with sharpe_ci_low -1.61 (CI straddles 0), and the thin +9.7%/5yr total is flattered by a Nov-2025 outlier (+8.7% day, kurtosis 7.84). No parameter change (volume_mult, volume_window, sma_trend_period, per_leg_pct) can recover a near-fee-fragile edge swamped by 27.6% impact and capped at ~$1.3M capacity -- these are consequences of ICP's low liquidity. Not iterate (code is correct and matches the hypothesis; the instrument is the problem). The volume-breakout family's value is already captured by the liquid majors; transplanting it to thin alts (TRX, ICP) yields a fee/impact-dominated non-edge deployable only at toy scale. FAILURE PATTERN: the FourHourVolumeBreakoutLong family does not generalize to low-liquidity alts -- on ICP it produces Sharpe 0.39, PF 1.07, impact_cost_pct 27.6%, and capacity ~$1.3M, i.e. the gross edge is largely consumed by market impact and the strategy is deployable only at toy scale. Liquidity, not signal, is the binding constraint; restrict this family to liquid majors.

Long-only single-instrument volume-breakout strategy on ICPUSDT.BINANCE USD-M futures 4H bars, using only OHLCV. It enters long when three conditions hold simultaneously on a closed 4H bar: (1) current volume exceeds 1.5x the prior 20-bar average volume, (2) the bar closes bullish (close > open), and (3) close is above its SMA(100) uptrend filter. It exits on a +5% take-profit, a -5% stop-loss, or a 14-bar (~2.3 day) max-hold time stop. calculate_signal returns a continuous graded score (tanh of the volume-surge ratio and bar body, modulated by the trend regime) that varies every bar and only reaches the >=0.5 entry band when all three conditions are met. Sizes 20% of equity notional per trade (ICP qty rounded to 0.1 step). Direct analog of the landed Btc/Eth/Sol/Bnb/Avax/Xrp/Matic/Fil FourHourVolumeBreakoutLong family.

Implemented correctly and trades as specified: 235 long-only volume-breakout entries, clean entry_diag (235 signaled = 235 submitted, 0 dropped), metrics_reliable=true; from a validated family

Catastrophic illiquidity: impact_cost_pct 27.6% (>25% threshold, impact eats a quarter+ of gross PnL) and capacity_usd only $1.3M (toy scale) -- a structural promotion blocker

Fee-fragile and weak: profit_factor 1.072 (1.0-1.1 zone), Sharpe 0.39, sharpe_ci_low -1.61 (CI straddles 0 widely), probabilistic_sharpe 0.646

Thin/choppy edge: total_return +9.7% over 5yr, negative 2021/2025, with a large Nov-2025 outlier (+8.7% day, kurtosis 7.84) flattering the tail

deployable scale

~0.5+ (pre-opt)

~62 WF floor

IcpFourHourVolumeBreakoutLong4H was another mechanical transplant of a promoted strategy family onto a thin alt, ICP. The code traded faithfully across 235 entries, but the result echoed the earlier TRX abandonment: the edge was both weak (profit_factor 1.072, Sharpe 0.39, a +9.7%/5yr return leaning on one Nov-2025 outlier) and structurally untradeable — market impact ate 27.6% of gross PnL and capacity topped out at $1.3M. The analyst abandoned it at backtest review without optimizing, since no parameter recovers an edge swamped by ~28% impact at toy scale. Together with TRX and DOT, ICP confirms the family's liquidity boundary: the volume-breakout edge is real and already captured on the liquid majors, and porting it to low-liquidity alts produces a fee/impact-dominated non-edge.

The FourHourVolumeBreakoutLong family does not generalize to low-liquidity alts — on ICP (like TRX) it yields Sharpe ~0.4, PF ~1.07, impact_cost_pct 27.6%, and ~$1.3M capacity, with the gross edge largely consumed by market impact and deployable only at toy scale — so a validated mechanism is bounded by its instrument's liquidity and this family should be restricted to liquid majors rather than ported further down the liquidity curve.

The analyst issued an abandon verdict at backtest review on a structural-liquidity basis identical to the earlier TRX arm: market impact consumes over a quarter of gross PnL (27.6%) and the strategy is deployable only at toy scale ($1.3M capacity), neither of which any parameter can fix — and the underlying edge is fee-fragile and weak (PF 1.07, Sharpe 0.39) regardless, so the instrument's low liquidity, not the implementation, is the binding constraint.

A long-only, OHLCV-only volume-breakout on ICPUSDT.BINANCE 4H futures — a direct transplant of the landed FourHourVolumeBreakoutLong family (BTC/ETH/SOL/BNB/AVAX/XRP/MATIC/FIL) — entering long on a three-condition conjunction (bar volume > 1.5× the prior 20-bar average AND a bullish bar AND price above its 100-SMA), exiting on a +5% take-profit, −5% stop, or 14-bar (~2.3-day) time stop.

It implemented and traded correctly with an ample sample — 235 long entries, clean entry diagnostics, metrics_reliable=true — but was structurally illiquid and only weakly positive: impact_cost_pct 27.6% (above the >25% red-flag) and capacity_usd just $1.3M, with profit_factor 1.072, Sharpe 0.39, sharpe_ci_low −1.61 (CI straddling zero widely), probabilistic_sharpe 0.646, and a thin +9.7% total return over 5 years that was negative in 2021/2025 and flattered by a Nov-2025 outlier (+8.7% day, kurtosis 7.84).