a1-return-autocorrelation-regime

a1 / a1-return-autocorrelation-regime @ v1 · code bff0511

campaign strategyREJECTED

Execution+Statistical Gate: Bootstrap 95% CI [-0.0047, 0.0123]R includes zero -- the raw edge is not statistically distinguishable from noise at this sample size.

Lifecycle

completed

Outcome

fail

Phase

Phase A

Research gate

EXECUTION_STAT

Engine

strategy lab

Market

MNQ

Timeframe

MNQ 1-minute OHLC

Side

BOTH

Tracking

Family, lineage and what this record needs next

Mechanism family

Regime-conditioned

a1-return-autocorrelation-regime

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 (A1)

research-output/strategy-lab/campaigns/a1/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

passed

Exec/stat

failed

Validation

—

H-eligible

—

Holdout

—

Audit

—

Closed

failed

INTAKE

Registered into campaign A1 at code version bff0511.

PREREGISTRATION

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

LIGHTNING

SURVIVED_LIGHTNING_SCREENAggregate expectancy 0.0044R (PF=1.007) over 136682 resolved trades, 4/5 qualifying years (n>=15) agree with the aggregate sign, top-10 winner concentration 3.4% (below the 50% concentration flag). Survives the Lightning Screen -- NOT validated.

ROBUSTNESS

ROBUSTNESS_SURVIVEDNo single-year dependence, no severe winner concentration (no preregistered neighborhood declared).

EXECUTION STAT

EXECUTION_STAT_FAIL_CI_INCLUDES_ZEROBootstrap 95% CI [-0.0047, 0.0123]R includes zero -- the raw edge is not statistically distinguishable from noise at this sample size.

CLOSED

REJECTEDExecution+Statistical Gate: Bootstrap 95% CI [-0.0047, 0.0123]R includes zero -- the raw edge is not statistically distinguishable from noise at this sample size.

Evidence recorded

Quoted from the campaign record — never recomputed here

Screen n (FIXED_2R)

136,682

Expectancy

+0.004R

Profit factor

1.01

Win rate

33.6%

Total R

597

Max DD

402R

Longest L streak

Trades / year

31864.0

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

Bootstrap CI

95% CI -0.005R .. +0.012R · P(expectancy > 0) = 80%

Winner concentration

Top 5 1.7% · top 10 3.4% · top 20 6.7% of total R.

Leave-one-year-out

No sign flip when any single year is removed.

Read from research-output/strategy-lab/campaigns/a1/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

Frozen specification

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

Mechanism hypothesis

Short-horizon return structure contains temporary momentum/reversion states measurable through recent autocorrelation.

Long rule

On a closed 5-minute bar i, compute lag-1 autocorrelation of closed-price log returns over the 30 closed 5-minute bars strictly before i. POSITIVE regime (autocorrelation >= 0.15): enter LONG (continuation) if bar i's own return is positive. NEGATIVE regime (autocorrelation <= -0.15): enter LONG (reversal) if bar i's own return is negative.

Short rule

POSITIVE regime: enter SHORT (continuation) if bar i's own return is negative. NEGATIVE regime: enter SHORT (reversal) if bar i's own return is positive.

Entry trigger

Evaluated once per newly-closed 5-minute bar: compute the 30-bar trailing lag-1 return autocorrelation (using only bars strictly before bar i), classify into POSITIVE/NEGATIVE/NEUTRAL, then apply the frozen state-to-direction mapping above to bar i's own just-closed return. NEUTRAL (autocorrelation strictly between the two thresholds) never trades.

Earliest executable entry

Next 1-minute bar's open after the triggering 5-minute bar closes.

Initial stop

1.5x ATR(14, 5-minute, computed only from closed bars strictly before the trigger bar) placed on the adverse side of the entry price -- a fixed, pre-registered volatility multiple, not an optimized or discretionary level.

Exit

FIXED_2R only (this seed strategy's declared A1 scope), via the shared TradeTracker; 5-day max-hold expiry as a backstop.

Causality

The autocorrelation value at bar i is computed only from closed 5-minute bars strictly before bar i; the return applied to the direction mapping is bar i's own just-closed return (known the instant bar i closes), and the ATR used for the stop is likewise computed only from bars strictly before bar i. Entry is deferred to the next 1-minute bar's open after bar i closes. No same-bar entry.

Expected frequency

MODERATE

Required data

MNQ 1-minute OHLC

Campaign data window

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

Parameters

autocorrWindowBars5m = 30 (integer) — 5-minute bars over which lag-1 return autocorrelation is computed.positiveAcThreshold = 0.15 (number) — Autocorrelation at or above this value is the POSITIVE (continuation) regime.negativeAcThreshold = -0.15 (number) — Autocorrelation at or below this value is the NEGATIVE (reversal) regime.atrLength5m = 14 (integer) — 5-minute ATR lookback used for the fixed-multiple stop.stopAtrMultiple = 1.5 (number) — Stop distance as a multiple of the 5-minute ATR.

Read from research-output/strategy-lab/campaigns/a1/campaign_state.json

History

Every dated event this record appears in, newest first

2026-09-15

a1-return-autocorrelation-regime@a1-v1 registeredREJECTED · EXECUTION_STAT

[a1] Execution+Statistical Gate: Bootstrap 95% CI [-0.0047, 0.0123]R includes zero -- the raw edge is not statistically distinguishable from noise at this sample size.

research-output/strategy-lab/experiment_registry.json