EthSessionOpeningRangeBreakoutContinuationLS1H

ETH Intraday Opening-Range Breakout Continuation, Long-Short (Binance USD-M, 1H, Session-Anchored at 00:00 UTC, Flat End-of-Day, ATR-Buffered Entry, Discrete Capital-Capped)

A LONG-SHORT, single-instrument, pure-OHLCV INTRADAY SESSION-MOMENTUM continuation strategy on ETHUSDT.BINANCE USD-M perpetual using 1-HOUR bars. Each UTC day it anchors an OPENING RANGE over the first few hours (from the 00:00 UTC daily/funding-clock boundary), then enters in the direction of the FIRST confirmed breakout of that range and rides the intraday continuation, flattening by end of day. This is a GENUINELY NEW mechanism (the analyst's explicit directive after the impulse-continuation family mapped its boundaries: promoted on ETH/SOL, FAILED on BTC/LINK/XRP/HL), distinct from every dead/blocked path: it is NOT a single-large-impulse-bar trigger (that family is saturated/asset-specific), NOT a daily-trend or relative-momentum majors-basket (all correlated re-expressions of the promoted AbsoluteMomentum, dead), NOT reversion (dead), NOT cross-venue funding carry (fee-squeezed), NOT options (engine-blocked), NOT sub-1H (fee-dominated). It is ALSO distinct from the regime-broken daily Donchian breakout BASKET: that was multi-instrument, multi-day trend-initiation whipsawed by daily chop; this is single-asset, intraday, anchored to a fixed session reference, and FLAT every night — so it captures intraday session-momentum (documented Asia/US-session directional drives and the daily-funding-clock effect) and resets daily rather than riding multi-day trends. It deliberately keeps the SAME validated core fact the impulse family exploits — crypto exhibits short-horizon CONTINUATION — but via an orthogonal, time-anchored trigger, on the asset where intraday continuation is most robustly proven (ETH, the promoted impulse asset), to maximize the edge prior. It is fee-survivable where sub-1H died (<=1 trade/day, flat EOD -> ~250 trades/yr, low fee incidence; a confirmed session break typically runs well past the ~0.10% RT) and validatable where HL intraday died (Binance USD-M ETH has deep 1H history -> ~1500 trades over ~6yr to populate the walk-forward). Engine-safe: discrete enter-once/exit-once, single position, hard-capped notional, no continuous rebalance/vol-scaling. Three core parameters.

Implements the hypothesis precisely: a time-anchored opening-range break is an ORTHOGONAL trigger to the single-impulse-bar continuation family (it captures Asia->US session directional drives and the daily-funding-clock effect rather than a magnitude-outlier bar), while keeping the SAME validated core fact (crypto exhibits short-horizon CONTINUATION) on ETH, the asset where intraday continuation is most robustly proven (promoted impulse asset). Flat-every-night + daily reset isolates intraday session momentum and avoids the multi-day daily-chop whipsaw that broke the Donchian breakout basket. Fee-survivable: <=1 trade/day flat EOD (~250 trades/yr, low fee incidence) and a confirmed session break typically runs well past the ~0.10% round-trip cost; deep Binance USD-M ETH 1H history (~6yr) populates the walk-forward (~1500 trades). Three core parameters (opening_range_hours, breakout_atr_buffer, stop_atr_mult) keep the surface small and overfit-resistant. Engine-safe and venue-correct: long+short directional perp strategy belongs on a futures venue (BINANCE USD-M); leverage left at 1.0 (sizing is risk-based + hard-capped, never exceeding equity notional).

No edge to optimize: the base backtest has profit_factor 0.997 (<1.0), win_rate 0.454, and negative expectancy (~-$4/trade with avg_win $2416 vs avg_loss $2016) — the opening-range breakout-continuation trigger is a coin-flip on ETH 1H. Every risk-adjusted metric is negative (Sharpe -0.354, Sortino -0.685, Calmar -0.10, IR -0.54), and the +169.88% headline return is entirely unrealized mark-to-market (end_unrealized_pct 176.75 > total_return, so realized PnL is negative). Compounded by a fatal capacity profile: avg_position 100% of equity, turnover 640x, impact_cost_pct 104% (impact exceeds gross PnL), capacity ~$92k. Optimization fits parameters but cannot create an edge that the base signal lacks — fixing the oversized sizing only drags the result toward break-even, not into positive expectancy. Annual returns (+46/+19/-28/+17/-49/+29/-2) are regime-dependent directional beta, consistent with this session's repeated breakout/Donchian-family failures; the 'ETH is the proven continuation asset' prior does not rescue a coin-flip trigger. Not worth the 2-hour optimization spend. FAILURE PATTERN: time-anchored opening-range breakout continuation (long-short, intraday, flat-EOD) on ETH 1H produces profit_factor ~1.0 / negative-Sharpe coin-flip with no gross edge, plus full-notional sizing that makes impact cost exceed gross PnL — the breakout-continuation family remains regime-dependent bull beta with no stable intraday edge even on the family's promoted asset.

Long-short, single-instrument, pure-OHLCV INTRADAY opening-range breakout continuation on ETHUSDT.BINANCE USD-M perp, 1-HOUR bars. Each UTC day, anchored at the 00:00 UTC boundary, it builds an opening range over the first `opening_range_hours` hours (rolling max-high/min-low), then enters ONCE in the direction of the first confirmed breakout (close beyond the OR edge by `breakout_atr_buffer` * ATR -> BUY above, SELL below) and rides the intraday continuation. Exits on an ATR stop (`stop_atr_mult` * ATR adverse from entry), the end-of-day flatten hour, or a new UTC day (hard flatten) -- so it is FLAT every night and resets daily. ATR-risk-based sizing (risk_per_trade_pct of equity at the stop) with a hard notional cap (max_notional_frac of equity); no leverage, no vol-scaling, single discrete enter-once/exit-once position.

Mechanism correctly implemented: ~1 trade/day, flat EOD, balanced long/short (1103/1026), clean entry diagnostics with zero drops

Trade count (2129 over ~6yr) is statistically ample and appropriate for the session-anchored design

No edge: profit_factor 0.997 (<1.0), win_rate 0.454, expectancy ~-$4/trade — a coin-flip with fatter losers

Every risk-adjusted metric negative: Sharpe -0.354, Sortino -0.685, Calmar -0.10, information_ratio -0.54, recovery_factor -0.10

Headline +169.88% return is unrealized mark-to-market (end_unrealized_pct 176.75 > total_return); realized PnL is negative

Catastrophic capacity/impact: avg_position 100% of equity, turnover 640x, impact_cost_pct 104% (exceeds gross PnL), capacity only ~$92k

Annual returns swing +46/+19/-28/+17/-49/+29/-2 — regime-dependent directional beta, not a stable intraday-continuation edge

EthSessionOpeningRangeBreakoutContinuationLS1H applied a 00:00-UTC-anchored opening-range breakout to ETH 1H perps, betting that ETH's validated intraday-continuation character would carry to an orthogonal, flat-EOD session-momentum trigger. The dense 2,129-trade sample made the verdict clear: no edge — profit factor 0.997, expectancy ~-$3.2, Sharpe -0.35 — with the eye-catching +169.9% return exposed as entirely unrealized mark-to-market over negative realized PnL, and a capacity wall where impact cost exceeded gross PnL. The analyst ruled it regime-dependent bull beta consistent with the session's repeated breakout/Donchian failures, not worth optimizing since tuning cannot create an absent edge. It ended after one iteration as abandoned, never advancing to optimization or risk review.

The 'ETH is the proven continuation asset' prior does not rescue a coin-flip trigger: the time-anchored opening-range breakout produced PF ~1.0 / negative-Sharpe regime-dependent bull beta even on the family's promoted asset, and a flattering headline return that is entirely unrealized mark-to-market masks negative realized PnL — confirming the breakout-continuation family has no stable intraday edge (it failed identically on SOL).

It was abandoned at the pre-optimization backtest-review gate (verdict: abandon) because the trigger is a coin-flip with no gross edge (PF ~1.0, negative expectancy and Sharpe) over an ample 2,129-trade sample that optimization cannot rescue, compounded by full-notional sizing whose impact cost exceeds gross PnL — so optimization and all later stages were never reached.

A single-instrument, pure-OHLCV long/short intraday opening-range breakout continuation strategy on ETHUSDT.BINANCE USD-M 1H perps — anchoring an opening range over the first few hours from 00:00 UTC, entering the first ATR-buffered breakout of that range, riding the intraday session-momentum, and flattening end-of-day — an orthogonal time-anchored continuation trigger applied to ETH, the family's promoted continuation asset, to maximize the edge prior.

Over ~6 years and a dense 2,129 trades (1,103 long / 1,026 short) it had no edge: profit factor 0.997, win rate 45%, expectancy ~-$3.2/trade, Sharpe -0.35 (CI [-1.08, 0.31]), Sortino -0.69, Calmar -0.10, 34.1% max drawdown; the +169.9% headline return was entirely unrealized mark-to-market (end_unrealized_pct 176.75 > total return, so realized PnL is negative), with a fatal capacity profile — avg position 100% of equity, turnover 640x, impact cost 104% of gross PnL, capacity ~$92k — and regime-swinging annual returns.