XrpFailedBreakoutSpringUpthrustReversalLS

Failed-Breakout (Spring/Upthrust) Reversal, Long-Short (Single-Instrument XRPUSDT.BINANCE USD-M Perp, 1h Bars, Fade a FALSE Break of a Multi-Bar Range — Enter Only When Price Pierces the Range Extreme Then Closes Back INSIDE = Trapped Breakout Traders, Tight Stop Beyond the Wick / Wide Target to Range Midpoint, 3-Parameter)

A LONG-SHORT, single-instrument, EVENT-DRIVEN reversal on XRPUSDT.BINANCE (USD-M perp, 1h bars). It trades the classic 'failed breakout' (Wyckoff spring / upthrust): when price briefly pierces a well-established N-bar range high or low and then CLOSES back inside the range on the same bar, the breakout attempt has failed and the traders who chased it are now trapped with stops sitting just beyond the extreme — that stop-run fuels a mechanical reversion back into the range. This is deliberately NOT another indicator fade (RSI/Bollinger/Keltner/VWAP z-score), the reversion class that died repeatedly this session with an INVERTED reward:risk (tiny take-profit + wide ATR stop => wins small, loses big). Here the geometry is inverted on purpose: the stop sits just beyond the rejected wick (small R) and the target is the range MIDPOINT (large relative to R), so the strategy is structurally reward-favorable. It also fires only on a discrete, rare trap event (a handful of setups per week), not on every moderate deviation, so cumulative fee drag stays low. Uses ONLY OHLCV — no supplementary feed that could dry up (the trap that killed the positioning/leverage-cycle baskets whose metric data ended in 2021).

Implements the failed-breakout trap: a pierce of a well-established N-bar range extreme that closes back inside traps the breakout chasers, whose stops just beyond the extreme fuel a mechanical reversion to the range center. It is deliberately NOT an indicator z-score fade (the RSI/Bollinger/Keltner/VWAP class that died with inverted reward:risk); the geometry is inverted on purpose — the stop sits just beyond the rejected wick (small R) and the target is the range midpoint (large relative to R), so the strategy is structurally reward-favorable and profitable at a modest hit rate, and it fires only on a discrete trap event so fee drag stays low. It uses ONLY OHLCV, avoiding the supplementary-feed data walls (metric series ending in 2021) that killed the positioning/leverage-cycle baskets. Risk control uses the pattern proven on the sibling perp strategies: entry fires only while flat (base guard plus explicit check, no pyramiding), and position_size bounds TOTAL gross exposure (net signed notional + new order) to max_notional_mult x equity = 0.5x under leverage 1.0, so a stop-out costs ~1% of equity and cannot liquidate. Single-instrument uses the base on_bar/dispatch unchanged; calculate_signal returns a continuous position-in-range score that varies every bar while caching the spring/upthrust booleans and the entry-anchored bracket (tight wick stop + midpoint target), and should_exit runs that bracket seeded via the _was_in_pos pattern. XRP is liquid with frequent range behavior on 1h, so failed-breakout setups recur several times per week -> hundreds of trades over the window (measurable), and a midpoint target is roughly half the range — on 1h XRP typically ~0.5-1.5%, clearing the ~0.10% USD-M round-trip several times over. Kept to exactly 3 tuned parameters (range_len, pierce_min, stop_buffer) to resist overfitting; the midpoint target, time cap, and risk caps are fixed structural choices. Both trap directions are mutually exclusive per bar (elif), and guards require the close on the correct side of the midpoint so the target is always reward-positive." }

Account liquidated (100% drawdown, -100% total return) on 2022-02-10 with negative per-trade expectancy across a decisive 2424-trade sample (avg_trade_return_pct -0.101%, PF 0.69, Sharpe -0.40). The failed-breakout thesis is falsified by its own realized numbers: despite a healthy 46% win rate, avg_win $243 is SMALLER than avg_loss $299 — the intended tight-stop/wide-midpoint-target geometry inverted in practice (the midpoint was rarely reached before the wick stop), landing in the same inverted reward:risk that killed the sibling single-name reversal strategies this session. Sizing also failed: avg_position_pct 104.9% shows full+ exposure despite the 0.5x notional cap, and catastrophic 2022 tail days (largest_loss -$3,837, kurtosis 748) blew up the account. Optimizing range_len/pierce_min/stop_buffer cannot flip negative expectancy or un-liquidate the account; it would only overfit the pre-blow-up window. Abandon per the >50% DD hard-abandon rule and the negative avg_trade_return_pct fee floor rather than spend 2 hours optimizing.

Long-short event-driven failed-breakout (Wyckoff spring/upthrust) reversal on XRPUSDT.BINANCE (USD-M perp), 1h bars, pure OHLCV. It defines the range over the prior range_len=20 bars (excluding the current bar). UPTHRUST -> SHORT: the bar's high pierces >= pierce_min (0.1%) above the range high but closes back inside (and above the range midpoint); SPRING -> LONG: the low pierces below the range low but closes back inside (and below midpoint). Each trade is bracketed with a tight stop just beyond the rejected wick (stop_buffer=0.3% beyond the extreme) and a target at the range MIDPOINT — small risk, large reward — plus a 24-bar time cap. Entry only while flat (no pyramiding); sizing risks 1% of equity over the tight stop and is bounded so total gross <= 0.5x equity (leverage 1.0, margin buffer). 3 tuned params (range_len, pierce_min, stop_buffer).

Abandon at backtest_review; re-tuning the 3 params can't fix the design flaw below.

STRONG ABANDON RECOMMENDATION (not a QA code fail). Decisively negative and well-sampled: 386 trades, total_return -31.98%, Sharpe -2.01 CI entirely below zero [-3.39, -0.64], PF 0.81, win_rate 0.43, avg_trade_return_pct 0.031% (below fee floor), turnover 248 / $10,866 impact, skew -4.74. Genuine edge failure, no degenerate signature — analyst's abandon call.

Flawed premise, not tuning — abandon. Would need a reward/R gate plus a real intrabar stop order (a redesign).

The hypothesis's central 'small R, large reward' claim is NOT enforced. Target = range midpoint, but entry only requires close on the far side of the midpoint — so a close just past mid yields a tiny reward while a tall rejection wick makes R large; the biggest traps get the WORST reward:risk. Sandbox confirms: avg_win $834 vs avg_loss $775 (1.08, not favorable) at 43% hit → negative expectancy, reproducing the inverted-RR failure the hypothesis claimed to avoid. Close-based exits also let losses blow through the tight wick-stop (skew -4.74). Code faithfully implements the described levels (not a critical structural mismatch), but the premise doesn't hold.

Abandon the pure-OHLCV framing; if iterated, gate the trap block on _open_position() is None.

Pure-OHLCV range-fade (0% survival across 213 runs). Minor latent smell: the trap/_last_trap write isn't flat-gated (unlike the fixed BNB sibling), so the bar after entry could overwrite _last_trap before _track_position seeds the bracket — non-catastrophic here.

Well-sampled (2424 trades, 57k bars) so the negative result is statistically decisive

Trade direction matches the hypothesis (balanced long-short: 1242 long / 1182 short)

Uses only OHLCV (no supplementary feed that dries up)

Account LIQUIDATED 2022-02-10: total_return -100%, max_drawdown 100% — hard-abandon (>50% DD)

avg_trade_return_pct -0.101% — negative per-trade expectancy below the fee floor

profit_factor 0.69, Sharpe -0.40, omega 0.67 — no edge

Core thesis falsified: despite a 46% win rate, realized avg_win $243 < avg_loss $299, so losses are BIGGER than wins — the exact inverted reward:risk the hypothesis claimed to avoid (the midpoint target was rarely reached before the wick stop)

avg_position_pct 104.9% — ran full+ exposure despite the 0.5x notional cap; catastrophic 2022 tail days (largest_loss -$3,837, kurtosis 748) drove the liquidation

~4 years of the backtest are dead (all-zero returns after the blow-up), so there is no post-2022 sample to optimize over

abandoned

This strategy set out to fade failed range breakouts on XRP, buying springs and shorting upthrusts with a tight wick stop and a wide range-midpoint target — deliberately inverting the tiny-TP/wide-stop geometry that had killed sibling reversal strategies this session. Over a decisive 2,424-trade sample the intended geometry inverted anyway: despite a 46% win rate, the midpoint target was seldom reached before the wick stop, so average losses ($299) exceeded average wins ($243), and the account was liquidated on 2022-02-10 for a -100% return and 100% drawdown. Compounding the failure, sizing ran at 104.9% average position despite a 0.5x notional cap, and extreme 2022 tail days (kurtosis 748) drove the blow-up. The analyst abandoned it at backtest review under the hard drawdown rule and negative fee-floor expectancy, noting optimization could neither flip the negative edge nor un-liquidate the account.

A geometrically 'reward-favorable' bracket on paper can invert in practice if the wide target is rarely reached before the tight stop, and a sizing model that lets exposure exceed its own notional cap (104.9% avg position vs a 0.5x cap) will blow up the account on tail days regardless of the entry logic.

The analyst abandoned it at the pre-optimization backtest-review gate under the >50% drawdown hard-abandon rule and the negative avg_trade_return_pct fee floor: the midpoint target was rarely reached before the wick stop (so losses ran bigger than wins despite a healthy hit rate), and sizing failed — avg_position_pct 104.9% ran full-plus exposure despite the 0.5x notional cap, with catastrophic 2022 tail days (largest loss -$3,837, kurtosis 748) driving the liquidation.

A long-short, single-instrument, event-driven reversal on XRPUSDT.BINANCE USD-M perps (1h bars, OHLCV only) that faded failed breakouts (Wyckoff spring/upthrust) — entering when price pierced a multi-bar range extreme then closed back inside, with a tight stop just beyond the rejected wick and a wide target at the range midpoint for a structurally reward-favorable geometry.

It was a decisively-sampled failure: the account was liquidated on 2022-02-10 with total return -100% and 100% max drawdown across 2,424 balanced trades (1,242 long / 1,182 short), win rate 46%, profit factor 0.69, Sharpe -0.40, and avg_trade_return_pct -0.101%. The reward-favorable thesis inverted in practice — realized avg win $243 was smaller than avg loss $299.