BTC 15-Minute Large-Impulse-Bar Continuation (Single-Instrument USD-M, Long-Short, Fast Intraday, Ride Directional Impulse Bars, Tight Initial Stop, Trailing-Stop Winners, Discrete Capital-Capped, Pure OHLCV)

A LONG-SHORT, SINGLE-INSTRUMENT, pure-OHLCV FAST-INTRADAY MOMENTUM-CONTINUATION strategy on BTCUSDT.BINANCE USD-M perpetual futures (15-MINUTE bars) -- the first TIMEFRAME extension of the factory's VALIDATED, analyst-endorsed IMPULSE-CONTINUATION family. The ETH 1H instance (EthImpulseBarContinuationTrendLS1H) PASSED optimization at Sharpe 1.40 and is PROMOTING; ETH/BTC/SOL/XRP/LINK 1H instances populate the family. The family is saturated at 1H (5 assets); its ONE unexplored axis is TIMEFRAME, which is also the biggest quota gap -- the short_1m_15m horizon bucket is 2.5% vs a >=10% target (the most under-represented). This applies the proven mechanism at 15m to fill that bucket. MECHANISM (mirrors the proven ETH template, faster grain): an IMPULSE bar is a 15m bar whose directional body (close - open) exceeds impulse_mult x ATR(atr_period) AND closes in the strong third of its high-low range -- a high-conviction directional thrust. The strategy enters in the impulse direction to capture short-horizon CONTINUATION, with a TIGHT initial stop and a TRAILING stop that lets winners run. WHY IMPULSE-CONTINUATION IS THE RIGHT VEHICLE FOR SUB-1H (and not the banned scalping anti-pattern): it is EVENT-DRIVEN and SELECTIVE -- a high impulse_mult means only RARE, LARGE 15m thrusts qualify, whose continuation move is large enough to clear fees; it does NOT trade every bar or chase sub-0.15% ticks. WHY 15m BTC: BTC is the DEEPEST, most-liquid major -> the cleanest 15m impulse signal (least microstructure noise) AND the largest continuation moves -> the best fee-clearing at the fast timeframe (the one real risk). WHY IT GENERALIZES WHERE OTHER SINGLE-ASSET INTRADAY DECAYED: impulse-continuation is event-driven/selective, the higher-quality signal that passed the ETH holdout, unlike VWAP-continuation (rode noisy excursions, decayed) and single-asset TSM/continuous-trend (whipsawed -- a DIFFERENT mechanism). WHAT IT AVOIDS (every closed path): NOT funding/carry/cross-venue (comprehensively closed -- uncreditable funding + HL-data wall); NOT a basket refinement (prohibited axis); NOT single-asset TSM/continuous-trend (dead); NOT reversion (dead); NOT continuous-rebalance/netting-flip (the 316%/385% sizing-artifact cause -- this uses DISCRETE enter-then-exit-to-FLAT, like the cleanly-sized ETH sibling). EVERY design choice respects the lessons: (1) DISCRETE entries, flat between signals, sized against STABLE realized equity (excl. unrealized PnL), hard-capped <=100% (leverage 1x). (2) FEE-AWARE BY CONSTRUCTION: a high impulse_mult restricts entries to large 15m thrusts whose continuation clears the ~0.10% RT floor -- NOT scalping; the backtest must confirm avg net trade > 0.15%. (3) DENSE: 15m gives the most impulse events of any family member -> robust 3-window walk-forward population. (4) PURE OHLCV. Per the 'favor simpler' mandate, 3 tunables (impulse_mult, trail_mult, init_stop_mult) mirroring the proven ETH config.

Extends the validated, promoting impulse-continuation family (ETH 1H sibling passed optimization at Sharpe 1.40) to the most under-represented horizon bucket (short_1m_15m, 2.5% vs >=10% target) by applying the identical mechanism at 15m on BTC, using the standard FactoryStrategy hooks. The impulse definition mirrors the proven template: directional body |close-open| >= impulse_mult*ATR AND close in the strong third of the high-low range (close-low or high-close >= 2/3 of range), a high-conviction thrust; entry is in the impulse direction for continuation. calculate_signal returns the continuous body/ATR so the signal varies every bar (Layer-2 frozen-signal detection passes) while the discrete impulse decision lives in should_enter. The exit combines a tight initial stop and a wider trailing stop into one rising floor (max(entry-init*ATR, peak-trail*ATR) for longs, mirrored for shorts): the tight initial stop cuts a failed impulse fast, and as a winner extends the trailing term overtakes it and lets continuation run -- no fixed take-profit. A max-hold cap (96 bars ~24h) is a safety backstop. It respects every lesson: DISCRETE enter-then-exit-to-flat (one position at a time, flat between signals -- not the continuous-rebalance/netting-flip that caused the 316%/385% sizing artifacts), sized against stable get_account_equity() (USDT balance excludes unrealized PnL) and hard-capped at max_notional_frac (95%) of equity at leverage 1.0 so exposure can never exceed ~1x; selectivity (high impulse_mult) restricts entries to rare large 15m thrusts whose continuation clears the ~0.10% RT fee, so it is fee-aware by construction and not scalping. Risk-based sizing (1% equity risk over the 1.0x-ATR initial stop) clears the BTCUSDT $100 min-notional. BTC is chosen as the deepest, least-microstructure-noisy major for the cleanest 15m impulse and the largest fee-clearing continuation. leverage stays 1.0, referenced only as a non-amplifying sizing multiplier, so the unused-leverage gate does not apply. Single-instrument pure OHLCV avoids the funding/cross-venue/tick/options walls; the analyzing stage should confirm avg net trade > 0.15% to validate fee-viability at the fast timeframe.

Third liquidation of the impulse-continuation family at the wrong grain (after XRP 1H and ETH 15m), and the clearest fee-death case: BTC 15m impulse continuation LIQUIDATED on 2021-11-02 (total_return -100%, max_drawdown 100%, dead 2022-2026) with a losing signal (profit_factor 0.763 < 1.0, Sharpe -0.29, win_rate 26%, losing every year). FEES are the assassin: the strategy fired 4305 trades at 336x turnover for a commission_pct_of_gross of 23.81% ($147k commission), and avg_trade_return is -$44 -- the average NET trade is negative, directly failing the hypothesis's own fee-viability gate ('the backtest must confirm avg net trade > 0.15%'). Both the 'fee-aware by construction' claim (high impulse_mult restricts to large thrusts that clear fees) and the 'BTC's large continuation moves clear fees at the fast grain' claim are falsified: BTC's depth does not change the arithmetic -- 15m impulse continuation cannot clear the ~0.10% round-trip at 336x turnover. The late-2021 chop whipsawed the continuation entries into a -100% spiral with physically-impossible prints (+292.9% on 2021-10-25, -150.5% on 2021-10-28, return_kurtosis 780). Notably the sizing was roughly within cap this time (avg_position 93.8% vs 0.95), so the failure is purely fee-death + losing signal, not the over-exposure bug. Not optimize: a series that liquidates with impossible bars, PF 0.763, and 24%-of-gross fees cannot be optimized, and the 3 tunables cannot lift a sub-fee edge. Not iterate: no sizing fix makes a 336x-turnover 15m strategy fee-viable, and the ETH 1H instance already captures the working version. FAILURE PATTERN: the impulse-bar-continuation mechanism is grain-bounded -- it works at 1H on ETH (Sharpe 1.40, promoted) but does NOT transfer to the 15m grain on ANY asset (ETH 15m AND BTC 15m both liquidated to fee death: commission 20-24% of gross, 229-336x turnover, avg net trade NEGATIVE), nor to other assets at 1H (XRP 1H liquidated). 'Proven asset x proven grain' is falsified because 15m is not a proven grain regardless of liquidity; sub-1H impulse continuation is fee-death territory, and a high impulse_mult does not make it fee-viable. The family is saturated at 1H on its proven assets; pushing it to faster grains reliably liquidates.