DotFourHourVolumeBreakoutLong4H

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

Long-only DOT (Polkadot) 4-hour strategy on BINANCE USD-M futures. Direct analog of the landed BTC/ETH/SOL/BNB/AVAX/XRP/MATIC/LINK FourHourVolumeBreakoutLong family — applied to DOT. 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.

Implements the hypothesis as a faithful analog of the proven FourHourVolumeBreakoutLong family: identical three-condition entry conjunction (volume surge + bullish bar + SMA trend filter), identical locked risk constants (+5%/-5%/14-bar), and identical capital-relative sizing, with only the instrument (DOTUSDT.BINANCE), class name, and quantity precision (DOT step 0.1 -> 1 decimal) changed. The 4H BINANCE USD-M data path is chosen specifically to avoid the daily-bar audit-refill failures that killed 12+ daily hypotheses this session, matching the hypothesis's pivot rationale. Routed to a futures venue (the family's data-path-proven home) but kept long-only with leverage 1.0 so sizing reads equity*per_leg_pct without margin amplification, sidestepping the leverage_set_but_unused gate. on_bar drives the real entry/exit logic deterministically while calculate_signal exposes a continuous tanh-squashed volume/body/regime score (varying every bar) so the verification layer sees graded conviction rather than a constant. The 1.5x volume surge plus 5% TP/SL band targets per-trade moves well above the ~0.10% round-trip taker cost.

Marginal edge on a lower-liquidity instrument that does not survive the validation gates -- exactly the abandon condition pre-registered at optimization. The optimizer improved DOT (Sharpe 0.51->1.69, PF 1.11->1.40, is_overfitted=FALSE with positive walk-forward OOS), but the decisive gates fail: deflated_sharpe 0.2392 (is_significant FALSE) with the optimized Sharpe 1.69 BELOW the expected-max luck bar 2.33 over 225 trials, PBO 0.664, sharpe_ci_low -0.2643 (straddles 0), sensitivity FAILED with 2 cliffs (incl. core volume_mult), and -- most decisively -- the HOLDOUT FAILED with a NEGATIVE Sharpe -1.54 (ratio -0.915): the once-tested recent window loses money, consistent with the recent decay (2025 -9.4%). The positive walk-forward OOS and is_overfitted=FALSE are real but outweighed by DSR 0.24 + negative holdout + sub-luck-bar Sharpe + PBO 0.66. Not iterate (attempt 1 of 2): DOT is the weakest, most fee-fragile arm of the volume-breakout family; the optimizer already found its best config and it still fails DSR, the holdout, and the luck bar -- a second sweep re-finds the same sub-luck-bar / 0-straddling / negative-holdout result. Not revise_hypothesis: the volume-breakout mechanism is already validated and promoted on the liquid majors (BTC/ETH/SOL/BNB/...), so its value is captured; DOT is simply the negative arm of the family's asset-selectivity, not a proven mechanism stranded on a dead target a different instrument would rescue. FAILURE PATTERN: the FourHourVolumeBreakoutLong family does not generalize robustly to lower-liquidity alts -- on DOT even the optimized config (Sharpe 1.69) fails deflated Sharpe (0.24, below the 2.33 luck bar), fails the holdout with a negative -1.54 Sharpe, and shows PBO 0.66 with core-param cliffs; the edge is real on the majors but too marginal/decaying on DOT to clear multiple-testing deflation and one-shot holdout validation.

Long-only volume-breakout strategy on DOTUSDT.BINANCE USD-M futures, 4-hour bars. Enters long when three conditions conjunct on a completed 4H bar: (1) bar volume exceeds volume_mult (1.5x default) the prior 20-bar average volume, (2) the bar is bullish (close > open), and (3) close is above the 100-period SMA uptrend filter. Sizes 20% of equity per position (leverage 1.0, no margin amplification). Exits on +5% take-profit, -5% stop-loss, or a 14-bar (~2.3 day) max-hold. Single-instrument, OHLCV-only — a direct port of the landed BTC/ETH/SOL/BNB/AVAX/XRP/MATIC/LINK FourHourVolumeBreakoutLong family applied to Polkadot, using DOT's 0.1 quantity step.

From the validated FourHourVolumeBreakout family (8 landed siblings); functions correctly with an ample sample (266 long entries, metrics_reliable=true)

Net positive across the period (total_return +25.5%) with slight positive payoff asymmetry (avg_win $1,422 > avg_loss $1,296) and contained drawdown (12.2%); positive in 4 of 7 years

Weakest arm of the family: Sharpe 0.51 (vs ETH 0.95 / BNB 1.52 / ADA 2.29), sharpe_ci_low -0.74 (CI straddles 0) -- strong DSR-fail risk after 225-trial selection

Fee-fragile: profit_factor 1.11 (in the 1.0-1.1 zone where costs eat most of the edge) combined with impact_cost_pct 14.3% and low capacity $4.86M (DOT illiquidity)

Front-loaded and lumpy: 2021 (+23.7%) dominates; 2022 (-8.8%) and 2025 (-8.9%) are large negative years

~0.5+ (pre-opt)

~62 WF floor

< ~25 (but high here)

Optimizer improved DOT meaningfully: Sharpe 0.51->1.69, PF 1.11->1.40, max_dd 9.65%; walk-forward is_overfitted=FALSE with OOS positive in all 3 windows ([4.40, 0.16, 0.49])

Functions correctly, market-neutral (beta 0.03), positive in 5 of 7 years in the optimized run

Fails deflated Sharpe: DSR 0.2392 (<<0.95), is_significant FALSE; optimized Sharpe 1.69 BELOW the expected-max luck bar 2.33 over 225 trials; PBO 0.664 (>0.5); sharpe_ci_low -0.2643 (straddles 0)

Holdout FAILED with a NEGATIVE Sharpe -1.54 (ratio -0.915) -- the once-tested recent window loses money (the decisive one-shot validation)

Sensitivity FAILED with 2 cliffs incl. the core volume_mult param; recent decay (2025 -9.4%, negative recent rolling Sharpe)

Weakest, most fee-fragile arm of the volume-breakout family; value already captured by promoted liquid-major siblings

> 2.33 (expected-max)

DotFourHourVolumeBreakoutLong4H transplanted a validated, promoted strategy family onto DOT, the lower-liquidity end of the asset spectrum. It behaved like the family's weakest arm from the start — net-positive (+25.5%) but with a sub-1.2 profit factor, a zero-straddling Sharpe CI, 14.3% impact cost, and a 2021-dominated curve — yet it cleared the review gate and optimization meaningfully improved it (Sharpe 0.51→1.69) while even passing the walk-forward overfit test. What killed it was the deflation layer: the optimized Sharpe fell below the 2.33 expected-max luck bar (DSR 0.24, PBO 0.66), and the once-tested holdout lost money at Sharpe −1.54, matching the strategy's recent decay. The analyst abandoned it on its first iteration — DOT is simply the negative arm of the family's asset-selectivity, its mechanism value already captured by the liquid majors, and a second optimization would only re-find the same sub-luck-bar, negative-holdout result.

The FourHourVolumeBreakoutLong family does not generalize robustly to lower-liquidity alts — on DOT even the optimized config (Sharpe 1.69) sits below the 2.33 luck bar, fails deflated Sharpe (0.24) and the negative holdout (−1.54), and shows PBO 0.66 with core-param cliffs — so a passing walk-forward is not enough when the holdout is negative and the optimized Sharpe doesn't clear the multiple-testing luck bar; the edge is real on majors but too marginal and fee-fragile on DOT.

The backtest-review gate passed it to optimization ('optimize'), but the analyst abandoned it post-optimization: despite a clean walk-forward, the decisive gates failed — deflated_sharpe 0.2392 (is_significant=false) with the optimized Sharpe 1.69 BELOW the expected-max luck bar of 2.33 over 225 trials, PBO 0.664, sharpe_ci_low −0.2643, 2 sensitivity cliffs (including the core volume_mult), and most decisively a holdout that FAILED with a negative Sharpe −1.54 (ratio −0.915) consistent with recent decay (2025 −9.4%).

A long-only, OHLCV-only volume-breakout on DOTUSDT.BINANCE 4H futures — a direct port of the landed FourHourVolumeBreakoutLong family (BTC/ETH/SOL/BNB/AVAX/XRP/MATIC/LINK) — 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.

The raw backtest was the family's weakest arm but net-positive — 266 long trades, total_return +25.5%, Sharpe 0.51 (CI straddling zero at −0.74 to 1.72), profit_factor 1.11, slight payoff asymmetry (avg_win $1,422 > avg_loss $1,296), max drawdown 12.2%, positive in 4 of 7 years — though fee-fragile (impact_cost_pct 14.3%, capacity $4.86M) and front-loaded to 2021. Optimization improved it to Sharpe 1.69 / PF 1.40 / total_return +39.5% and passed walk-forward (is_overfitted=FALSE, OOS [4.40, 0.16, 0.49]), but failed the deflation and one-shot gates.