a03-failed-breakout-reentry

a0 / a03-failed-breakout-reentry @ a0-regression-v1 · code 0c0b1d0

campaign strategyREJECTED

Robustness Gauntlet: Removing 2024 alone flips the combined expectancy sign -- the whole edge rests on one year (that year's own totalR=114.0).

Lifecycle

completed

Outcome

fail

Phase

Phase A

Research gate

ROBUSTNESS

Engine

strategy lab

Market

MNQ

Timeframe

MNQ 1-minute OHLC

Side

BOTH

Tracking

Family, lineage and what this record needs next

Mechanism family

Breakout / failed breakout

a03-failed-breakout-reentry

Mechanism parent

None declared.

Phase detail

Scientific validation lane (Phase A).

Next required test

None — this record is complete.

Integrity & certification

Signals as recorded; NOT_PASSED means not passed or not applicable

Engine certified

unknown

Execution certified

unknown

Integrity audit

unknown

Shadow verifier

unknown

Technical invalidation

None recorded.

Artifacts

Everything committed for this record

Campaign record (A0)

research-output/strategy-lab/campaigns/a0/campaign_state.json

Stage ladder

Each cell carries the campaign runner's own classification label — hover for the verdict and its reason

Intake

passed

Pre-reg

passed

Lightning

passed

Robust

failed

Exec/stat

—

Validation

—

H-eligible

—

Holdout

—

Audit

—

Closed

failed

INTAKE

Registered into campaign A0 at code version 0c0b1d0.

PREREGISTRATION

Pre-test passed: distinct, causal, executable, required data available.

LIGHTNING

SURVIVED_LIGHTNING_SCREENAggregate expectancy 0.0100R (PF=1.015) over 8706 resolved trades, 3/5 qualifying years (n>=15) agree with the aggregate sign, top-10 winner concentration 23.0% (below the 50% concentration flag). Survives the Lightning Screen -- NOT validated.

ROBUSTNESS

ROBUSTNESS_FAIL_ONE_YEAR_DEPENDENCERemoving 2024 alone flips the combined expectancy sign -- the whole edge rests on one year (that year's own totalR=114.0).

CLOSED

REJECTEDRobustness Gauntlet: Removing 2024 alone flips the combined expectancy sign -- the whole edge rests on one year (that year's own totalR=114.0).

Evidence recorded

Quoted from the campaign record — never recomputed here

Screen n (FIXED_2R)

8,706

Expectancy

+0.010R

Profit factor

1.02

Win rate

33.9%

Total R

Max DD

145R

Longest L streak

Trades / year

2030.0

Split	Window	n	Expectancy	PF	Max DD
Discovery	—	—	—	—	—
Validation	—	—	—	—	—
Holdout	—	—	—	—	—

Winner concentration

Top 5 11.5% · top 10 23.0% · top 20 46.0% of total R.

Leave-one-year-out

Sign flips when removing: 2024.

Read from research-output/strategy-lab/campaigns/a0/campaign_state.json

Holdout position

One-shot, sealed, never retried

Eligible

—

Authorised

Not authorised.

Specification frozen

—

Sealed data consumed

No.

Validation data consumed

No.

Lineage

Declared at intake, carrying any prior negative result

Relation to prior work

new independent mechanism

Descendants

a5-07-compression-break-false-failure a5-13-prior-day-extreme-failed-acceptance

Frozen specification

Snapshotted at intake — not re-read from a live module

Mechanism hypothesis

A meaningful recent range break that quickly fails and re-enters the prior range can indicate trapped breakout participation and short-term reversal pressure.

Long rule

If a closed 5-minute bar's close breaks below the trailing 12-bar (60-minute) range low by at least 0.2x ATR(14,5M) (the range computed only from bars strictly before the breakout bar), and within the next 3 closed 5-minute bars price closes back at/above the ORIGINAL range low: enter LONG at the next 1-minute bar's open.

Short rule

Symmetric: a break above the trailing range high by the same ATR-normalized distance, followed by a close back at/below the original range high within 3 bars, enters SHORT.

Entry trigger

Failed-breakout confirmation: a valid-distance range break followed by a close back inside the ORIGINAL (frozen-at-breakout) range within 3 closed 5-minute bars. At most one breakout attempt tracked at a time -- a new signal while one is already pending is skipped, not queued.

Earliest executable entry

Next 1-minute bar's open after the failure/re-entry-confirming 5-minute bar closes.

Initial stop

The breakout attempt's own most extreme point reached during the whole excursion (from the breakout bar through the failure-confirming bar) -- beyond this level the failed-breakout thesis is invalidated.

Exit

FIXED_2R only (this seed strategy's declared A0.3 scope), via the shared TradeTracker; 5-day max-hold expiry as a backstop.

Causality

The recent range (high/low) is computed only from the 5-minute bars STRICTLY PRECEDING the candidate breakout bar and is frozen at the instant the breakout bar closes -- never recomputed or extended using any later bar, so the range is completely known before the breakout occurs. Entry is deferred to the following 1-minute bar's open, never the confirmation bar itself. No same-bar entry.

Expected frequency

LOW

Required data

MNQ 1-minute OHLC

Campaign data window

MNQ 2022-06-03..2026-09-15

Parameters

rangeWindowBars5m = 12 (integer) — Trailing, strictly-preceding 5-minute bars defining the recent range.atrLength5m = 14 (integer) — ATR lookback in 5-minute bars, used for the minimum breakout distance.minBreakoutDistanceAtrMult = 0.2 (number) — A breakout must clear the range boundary by at least this multiple of ATR to be valid.failureHorizonBars5m = 3 (integer) — Number of closed 5-minute bars within which the breakout must fail and re-enter the original range, or the attempt expires unresolved.

Read from research-output/strategy-lab/campaigns/a0/campaign_state.json

History

Every dated event this record appears in, newest first

2026-09-15

a03-failed-breakout-reentry@a0-v1 registeredREJECTED · ROBUSTNESS

[a0] Robustness Gauntlet: Removing 2024 alone flips the combined expectancy sign -- the whole edge rests on one year (that year's own totalR=114.0).

research-output/strategy-lab/experiment_registry.json

2026-09-15

a03-failed-breakout-reentry@a04-robustness-v1 registeredREJECTED · ROBUSTNESS

Phase A0.4 Robustness Gauntlet -- FAIL. Bootstrap 95% CI [-0.0238, 0.0389]R comfortably includes zero -- the raw +0.0100R edge is not statistically distinguishable from noise at this sample size. Edge does not survive even 0.25pt adverse slippage. Sign flips when removing: 2024.

research-output/strategy-lab/experiment_registry.json

2026-09-15

a03-failed-breakout-reentry@v1 registeredSURVIVED_LIGHTNING · LIGHTNING_SCREEN

Phase A0.3 seed batch -- SURVIVED_LIGHTNING_SCREEN. Aggregate expectancy 0.0100R (PF=1.015) over 8706 resolved trades, 3/5 qualifying years (n>=15) agree with the aggregate sign, top-10 winner concentration 23.0% (below the 50% concentration flag). Survives the Lightning Screen -- NOT validated; Holdout/walk-forward/robustness work is explicitly out of scope for A0.3.

research-output/strategy-lab/experiment_registry.json