BtcDailyRangeExpansionClimaxReversalLongShort

BTC Single-Asset Range-Expansion Climax Reversal Long-Short on BINANCE Daily (Linda Raschke Climax Fade Pattern, Event-Driven, Translation-Robust)

Single-asset range-expansion climax reversal strategy on BTCUSDT.BINANCE — fades exhaustion bars where a single bar produces abnormally large range AND closes at an extreme of that range, indicating capitulation (sell-climax) or euphoria (buy-climax). Mechanism: (1) detect 'range expansion' bars where (high - low) > expansion_threshold × avg_range_20d (default 2.0×). (2) within an expansion bar, measure close_position = (close - low) / (high - low). (3) SHORT if expansion AND close_position > 0.75 (closed in upper 25% — buying climax exhaustion, fade up move). LONG if expansion AND close_position < 0.25 (closed in lower 25% — selling climax exhaustion, fade down move). FLAT otherwise. This implements Linda Raschke's 'Street Smarts' (1996) climax reversal pattern and Toby Crabel's 'Day Trading with Short Term Price Patterns' (1990) exhaustion fade methodology. The mechanism is FUNDAMENTALLY EVENT-DRIVEN: a 2×-normal-range bar with extreme close is a discrete event identifiable on ANY timeframe (the comparison is to the rolling average range at that same timeframe, so the criterion scales). This directly addresses the analyst's CRITICAL GENERALIZATION from the abandoned BTC TSM: low-frequency factors fail when ported to high-frequency, but EVENT-DRIVEN PATTERNS survive translation because the trigger semantics are invariant. Different mechanism class from all my in-pipeline single-asset strategies: TSM (vol-normalized t-stat — refuted), RSI MR (level signal), Macro Regime (slow MA crossover), Vol Compression Breakout (compression-then-direction-breakout). Climax Reversal is FADE-AT-EXHAUSTION, opposite direction to breakout strategies. Single-asset scope = no multi-asset translation curse. Helps direction quota (90.1% long-only — strategy goes both ways).

Implements Linda Raschke's 'Street Smarts' and Toby Crabel's exhaustion-fade climax reversal pattern. The edge is event-driven and scale-invariant: the trigger compares one bar's range to the rolling average range at the same timeframe, so the 2x-normal-range-with-extreme-close criterion has identical semantics on any interval, directly addressing the analyst's generalization concern that low-frequency factors fail under timeframe translation while event-driven patterns survive. It is a fade-at-exhaustion mechanism (opposite direction to breakout strategies) and genuinely two-sided, helping the long-only direction quota. Long+short requires a futures venue (BINANCE); risk-based sizing controls notional so no leverage is needed (leverage=1.0). The continuous signal (expansion_ratio x (close_pos-0.5)) is returned every bar for verifier visibility, while the faithful expansion>=2.0 AND close_pos in extreme-quartile conditions are checked in should_enter against stored state.

Negative-edge premise, not a coding or tuning problem. The strategy implements the Raschke/Crabel climax-fade exactly as hypothesized (75 trades, correct long/short polarity, trades match the mechanism), but fade-at-exhaustion has a structurally negative expectancy on BTC daily: PF 0.76, Sharpe -0.62, expectancy -$172/trade, and losses in 5 of 7 years across regimes. The signature failure is the COVID selling-climax (long into a -10.3% crash day) producing a -$10,334 outlier loss against a $1,305 avg loss (skew -4.47, kurtosis 61) — climaxes on a momentum-driven major continue rather than reverse, so systematically fading them loses. Optimization cannot raise PF 0.76 above the 1.2 viability bar; with a 75-trade sample showing a broadly negative parameter space, a sweep would only over-select a noise corner that fails walk-forward/holdout. Not iterate: the code is correct and the premise (not the parameters) is the flaw. FAILURE PATTERN: single-asset range-expansion climax-FADE long-short on a crypto major has negative edge because exhaustion bars on momentum-driven majors tend to continue, not reverse; the strategy is repeatedly run over by the very climaxes it fades (COVID -10.3% day = largest loss), yielding PF<1, negative Sharpe, and a fat negative tail that no threshold tuning can fix.

Single-asset long/short range-expansion climax reversal on BTCUSDT daily futures. Each bar it measures (high-low) versus the 20-bar average range to get an expansion ratio, and the close's position within the bar's range. When a bar's range exceeds 2.0x the average (a discrete range-expansion event) AND it closes in the upper 25% of its range, it SHORTs (fading a buying climax); when it closes in the lower 25%, it LONGs (fading a selling climax). Positions are risk-sized to 1% of equity against a 4% stop, exited on a 4% stop-loss, 6% take-profit, or after 6 bars. The signal is a continuous expansion-weighted deviation of the close from the bar midpoint, so it varies every bar; the two-part exhaustion trigger is applied in should_enter.

Code correctly implements the hypothesis: 75 event-driven entries (30 long / 45 short), trades match the fade-at-exhaustion mechanism with no polarity or sizing bug

Reasonable trade count (75) and both-directions exposure as the hypothesis intended

Low exposure (14.3%) and benchmark-independent (beta 0.005) — it is a genuine standalone signal, just a losing one

Fundamentally negative edge: profit_factor 0.76, Sharpe -0.62, expectancy -$172/trade, total_return -14.7%, CAGR -2.2%

Loses money in 5 of 7 years across multiple regimes — not a tunable/regime-specific issue

Severe negative tail: largest_loss -$10,334 vs avg_loss $1,305, skew -4.47, kurtosis 61 — fading the COVID selling-climax (long into a -10.3% crash day) is the mechanism failing exactly as designed

Premise flaw: range-expansion climaxes on BTC daily tend to CONTINUE (momentum), so systematically fading them has a structural negative expectancy

Optimizing cannot lift PF 0.76 past the 1.2 viability bar; it would only over-select a noise corner that fails walk-forward/holdout

This strategy faded range-expansion climaxes on BTC daily — shorting euphoric upper-extreme bars and longing capitulation lower-extreme bars — pitched as an event-driven pattern robust to timeframe translation. It implemented the Raschke/Crabel mechanism faithfully over 75 two-sided trades, but had a structurally negative edge: profit factor 0.76, Sharpe -0.62, -14.7% total return, losses in five of seven years, and a severe negative tail headlined by going long into the COVID -10.3% crash for its largest loss. The analyst abandoned it at the backtest-review gate on the first iteration as a flawed premise — climaxes on momentum-driven majors continue rather than reverse — not a tunable parameter problem, so it never advanced to optimization or risk review.

An event-driven pattern surviving translation does not make it profitable — fading exhaustion bars on a momentum-driven crypto major has structurally negative expectancy because those climaxes continue rather than reverse, so the strategy gets run over by the very moves it fades (e.g. longing the COVID crash), producing a fat negative tail no threshold tuning can fix.

The backtest-review analyst issued an 'abandon' verdict: the code correctly implemented the climax-fade mechanism, so this is a negative-edge premise rather than a bug or tuning issue — range-expansion climaxes on a momentum-driven major like BTC tend to continue, not reverse, so systematically fading them loses, and a profit factor of 0.76 cannot be lifted past the 1.2 viability bar by optimization.

A single-asset long-and-short range-expansion climax-reversal strategy on BTCUSDT daily Binance futures (ATR-style risk sizing) implementing the Raschke/Crabel exhaustion-fade pattern — shorting a buying climax (a 2x-normal-range bar closing in its upper 25%) and longing a selling climax (closing in its lower 25%) — pitched as an event-driven, translation-robust mechanism.

Across 75 trades (30 long, 45 short) it returned -14.7% with profit factor 0.76, Sharpe -0.62, expectancy -$172/trade, CAGR -2.2%, a 45.3% win rate and an 18.7% max drawdown, losing in 5 of 7 years; it carried a severe negative tail (largest loss -$10,334 vs avg loss $1,305, skew -4.47, kurtosis 61), exemplified by going long into the COVID -10.3% crash day.