Breakout

Definition

Breakout strategies enter in the direction of a move through a defined boundary — a prior high/low, a consolidation range edge, a channel line — on the premise that boundary breaks mark the start of continuation rather than the end of the move. The style trades late entry (paying confirmation) for trend participation, accepting that many breaks fail.

How it works / structure

• Boundary definitions: N-day high/low (Donchian channels — the classic trend-entry rule, indicator-donchian-channels), consolidation range edges, volume-profile shelf edges (indicator-volume-profile), or 52-week extremes (range_52w_pos_pct).
• Parameters (engine-executable): boundary lookback, entry style (stop-through vs close-through — intrabar breaks fail more often), confirmation filter (rel_volume above threshold, minimum ATR expansion), initial stop (back inside the range, or ATR-scaled — atr_14_pct), trail rule (mgmt-stop-loss trailing variant), position size from stop distance (risk-fixed-fractional).
• Failure-case economics: the strategy’s distribution is many small losses (failed breaks) against few large trend gains — payoff-skew, not hit-rate, carries it; accounting must judge it accordingly.

When it applies

Instruments and regimes where trends follow range resolution (volatility compression → expansion), futures trend systems (Donchian breaks are the entry engine of strategy-futures-trend-following), and event-driven range resolutions with the event filter stated. Thin books make breakouts self-defeating — the entry demands liquidity at the worst moment (ms-liquidity).

Risk profile & failure modes

• False-break rate is the cost center: boundary-adjacent stops are visible liquidity; sweeps through obvious levels that immediately reverse are a structural feature of auction markets, not bad luck.
• Slippage at the break: breakout entries are marketable orders in a fast market — realized entry prices degrade exactly when the signal fires (ms-slippage-friction).
• Chop destruction: range regimes whipsaw breakout systems repeatedly; the regime filter determines survival (regime-volatility).
• Data-snooping: boundary/lookback combinations are a large search space — Sullivan-Timmermann-White (1999) applies with full force.

Evidence & limits

Brock-Lakonishok-LeBaron (1992) found range-break rules among the tested rules with statistically significant predictive content on long Dow history; subsequent cost- and snooping-adjusted work weakened the net-profitability claim. Time-series momentum evidence (cited in strategy-momentum) is consistent with continuation after strength, which is the breakout’s premise at longer horizons. Specific chart-pattern breakout claims (flags, triangles, “measured moves”) are folklore absent citations.

Falsifiable-thesis examples

Illustrations only, not signals:

• “X, breaking its 55-day high on 2x relative volume, will reach +3 ATR before returning inside the prior range” — falsified by which level is hit first.
• “Donchian-55 entries on universe U with a 20-day trail will be net profitable after friction over the next two quarters in replay” — falsified by the replay P&L.

Cross-references

• Trend engines it feeds: strategy-futures-trend-following, strategy-momentum
• Boundaries and filters: indicator-donchian-channels, indicator-volume-profile, indicator-atr, rel_volume concept
• Execution reality: ms-liquidity, ms-slippage-friction, ms-sessions-auctions
• Management: mgmt-stop-loss (initial + trail), risk-fixed-fractional