BtcVolatilityExpansionMomentumContinuationLS

BTC Perp 1H Volatility-Expansion Momentum Continuation (Long-Short, Range-Expansion Triggered, ATR-Trailed)

A long-SHORT, single-instrument, single-venue, OHLCV-ONLY intraday momentum-continuation strategy on BTCUSDT.BINANCE perpetual futures using 1-HOUR bars. Core thesis is volatility clustering plus short-term momentum: a sudden RANGE EXPANSION bar with a decisive directional close reflects an aggressive order-flow impulse (a large market-order sweep, a liquidation cascade, or a genuine breakout) that tends to CONTINUE over the next several hours as leveraged positioning adjusts and trend-followers pile in, before the move eventually mean-reverts. The strategy trades the continuation in the direction of the expansion bar. It deliberately does NOT require a prior low-volatility compression window — that conjunction is exactly what made the previously-abandoned Bollinger-squeeze strategy too rare to populate the walk-forward (event-sparsity, all-zero OOS). By triggering on ANY sufficiently large directional bar, qualifying events occur several times per week, yielding hundreds of trades over BTC's ~6-year 1H history (>>10 entries per walk-forward OOS window and a populated holdout). It is symmetric long-AND-short, diversifying a portfolio that is 86% long-only, and uses only deep-history Binance OHLCV (no options/HL/COIN-M/cross-venue/supplementary-stream dependencies, all of which are confirmed unvalidatable in the current engine). Kept to 4 core parameters to resist the overfitting that produced is_overfitted=TRUE in this session's thin-conjunction strategies.

Implements the hypothesis directly: it triggers on ANY sufficiently large directional range-expansion bar rather than requiring a rare low-volatility compression conjunction, so qualifying events occur several times per week and produce hundreds of trades across BTC's multi-year 1H history — populating the walk-forward OOS windows and holdout that doomed the prior Bollinger-squeeze strategy. The signal is genuinely continuous (expansion ratio scaled by close location), positive for bullish impulses and negative for bearish, so the strategy is symmetric long-and-short, diversifying a long-heavy portfolio. It uses only deep-history Binance OHLCV with no options/HL/COIN-M/cross-venue/supplementary dependencies. Volatility clustering motivates ATR-relative expansion detection; short-term momentum motivates trading the continuation; the ATR trailing stop captures the trend while it persists and exits when momentum fades, with a max-hold cap honoring the eventual mean-reversion. Parameter surface is kept to four core knobs (atr_period, entry_threshold, trail_atr_mult, risk_per_trade) plus conservative supporting defaults to resist the overfitting seen in thin-conjunction strategies. Routed to futures (BINANCE) because the strategy goes short and uses 3x leverage, which position_size() actively consumes as a notional cap so the backtest is not inert.

BTC 1H vol-expansion momentum-continuation L/S is doubly broken: it LIQUIDATED on 2021-07-17 (max_drawdown 100%, total_return -100%, all daily returns 0.0 thereafter) AND its underlying signal has a negative/inverted edge. The continuation thesis is empirically wrong: PF 0.49, win_rate 0.35, expectancy -$57/trade over 2,507 trades, and EVERY month of 2020 (a full clean year before any leverage blow-up) is negative (-9% to -23%) — large 1H directional/vol-expansion bars on BTC MEAN-REVERT rather than continue, so trading WITH the impulse loses systematically. On top of the inverted edge, the position sizing over-leveraged to avg_position_pct 218% (the ATR risk-sizing consumes leverage against a 0.95 notional cap), producing impossible pre-liquidation daily returns (2021-07-04 -718%, 2021-07-01 +212%) as equity went negative and the account was liquidated; and the 1-hour-hold churn makes it fee-fragile (commission 57.27% of gross). Not optimize: a liquidated, PF-0.49 negative-edge strategy has nothing to optimize. Not iterate: unlike a pure sizing bug, fixing the leverage here only converts a fast liquidation into a slow bleed because the signal itself loses every month in a clean year; the momentum-continuation premise is inverted and the high-churn fee drag (57% of gross) is structural. FAILURE PATTERN: intraday (1H) momentum-CONTINUATION on a single crypto major trades WITH sharp directional bars that actually revert, producing a systematically negative edge (PF <0.5, every-month-negative), and when paired with ATR-risk sizing that consumes leverage it over-positions (~2x+) and liquidates on a volatility cluster (June-July 2021). This is the continuation-side mirror of the abandoned VWAP-excursion-MR family's inverted payoff: neither trading-with nor fading sharp intraday BTC bars has a fee-surviving edge, and ATR-risk sizing without a hard gross-exposure cap is a liquidation hazard. The Research Lead should retire single-asset intraday impulse-continuation strategies, and any future ATR-risk-sized strategy must hard-cap total gross exposure at <=1x equity.

Long-and-short, single-instrument intraday momentum-continuation strategy on BTCUSDT.BINANCE perpetual futures using 1-HOUR bars (OHLCV only, no supplementary data). Each bar it computes a continuous signal = expansion * CLV, where expansion = current bar's true range divided by the trailing ATR over the prior 14 bars (current bar excluded), and CLV = 2*(close-low)/(high-low)-1 measures how decisively the bar closed (+1 on the high, -1 on the low). A large positive signal marks an aggressive upside impulse (sweep/breakout/short-squeeze); a large negative signal an aggressive downside impulse (liquidation cascade). It enters in the direction of the impulse when |signal| >= entry_threshold (1.5), betting the move continues over the next several hours. Exits use an ATR trailing stop (3*ATR from the best price since entry), a catastrophic hard stop (2*ATR from entry), and a 48-bar (~2 day) max-hold cap. Positions are sized by risking 2% of equity over the 2*ATR stop distance, capped at 95% of equity * leverage (3x).

Genuinely high trade frequency (2,507 trades) solving the event-sparsity problem the hypothesis targeted; OHLCV-only, no data-coverage issue

Clean signal construction (expansion x CLV)

LIQUIDATED 2021-07-17: account wiped to zero (max_drawdown 100%, total_return -100%, all returns 0.0 after liquidation)

Inverted/negative edge: PF 0.49, win_rate 0.35, expectancy -$57/trade, and EVERY month of clean-year 2020 negative (-9% to -23%) — big 1H vol-expansion bars mean-revert, so trading the continuation loses

Catastrophic over-leverage: avg_position_pct 218%, impossible pre-blowup daily returns (-718%, +212%) as equity went negative

Fee-fragile: commission 57.27% of gross from 1-hour-hold churn (2,507 trades); annualized_vol 111.7%, kurtosis 539

BtcVolatilityExpansionMomentumContinuationLS bet that large 1H range-expansion bars on BTC signal continuation, trading symmetrically long and short with ATR-based stops and sizing, explicitly designed for high trade frequency to avoid the event-sparsity that killed prior squeeze strategies. It did fire often (2,507 trades) but the continuation thesis was empirically wrong: profit factor 0.49, 35% win rate, negative expectancy, and every single month of the clean 2020 year in the red, because such impulses mean-revert instead. Compounding the inverted edge, ATR-risk sizing pushed average position to 218% of equity and the account was liquidated outright on 2021-07-17 with a 100% drawdown, while 57% of gross was consumed by commissions. The analyst abandoned it at the backtest-review gate before any optimization or risk review, flagging it as the continuation-side mirror of the already-retired VWAP-excursion family and recommending the Research Lead retire single-asset intraday impulse strategies entirely.

Sharp intraday (1H) directional bars on a single crypto major tend to mean-revert rather than continue, so trading with the impulse has a fee-eroding negative edge, and any ATR-risk-sized strategy must hard-cap total gross exposure at ≤1x equity to avoid liquidation on volatility clusters.

The analyst issued an 'abandon' verdict at the pre-optimization backtest-review gate, so optimization, full analysis, and risk review were never reached: a liquidated, PF-0.49 negative-edge strategy has nothing to optimize, and fixing the leverage would only convert a fast liquidation into a slow bleed because the continuation signal itself loses systematically.

A long-and-short, single-instrument, OHLCV-only intraday momentum-continuation strategy on BTCUSDT.BINANCE 1H perpetual futures, entering in the direction of a large range-expansion bar (expansion × close-location-value signal) on the thesis that decisive volatility-cluster impulses continue for several hours, managed by ATR trailing/catastrophic stops and a max-hold cap.

It traded heavily (2,507 trades, solving its target event-sparsity problem) but had a clearly inverted/negative edge — profit factor 0.49, win rate 0.35, expectancy -$57/trade, Sharpe -0.75, with every month of clean-year 2020 negative (-9% to -23%) — and ATR-risk sizing over-leveraged to avg position 218%, driving a 100% drawdown that liquidated the account on 2021-07-17 (total return -100%); commissions were 57.27% of gross.