Fading QQQ Overnight Gaps

Conclusion: Rejected

Out-of-sample Sharpe was -0.62, versus -1.34 in-sample. A bootstrap test returned p = 0.774 (not significant).

Trade count is ample (407 OOS trades, well above the 30 minimum), so this is a genuine rejection rather than an inconclusive small-sample result. The documented positive overnight-drift effect does not imply mean-reversion of gaps intraday; if anything QQQ gaps trend.

Method

Rule / pseudocode

DATA ← adjusted observations for the stated instrument and horizon

RULE ← On QQQ, large overnight gaps partially reverse during the regular session, so fading the gap intraday (short up-gaps, long down-gaps, exit at close) earns a positive after-cost return.

METHOD ← Instrument: QQQ

FIT ← estimate or select parameters on each training window only

TEST ← apply the frozen rule to the next unseen window after stated costs

REPORT ← return, Sharpe, drawdown, trade count, significance and fold stability

VERDICT ← prefer robustness and sufficient observations over the headline return

_Overnight gaps are frequently driven by thin pre-market order flow and sentiment; RTH liquidity provision pushes price back, so extreme gaps revert more often than they continue._

Out-of-sample equity curve

Growth of $1 over the held-out test data the strategy never saw while it was being built, after 10 bps round-trip costs.

Out-of-sample performance

OOS Sharpe

-0.62

Ann. return

-14%

Max drawdown

-30%

Win rate

45%

p-value

0.774

not significant

WF folds +

2/5

Measured on held-out data. A high Sharpe with a non-significant p-value means the result could still be luck — read them together.

Hypothesis

On QQQ, large overnight gaps partially reverse during the regular session, so fading the gap intraday (short up-gaps, long down-gaps, exit at close) earns a positive after-cost return.

Method

Instrument: QQQ

Results

Out-of-sample Sharpe was -0.62, versus -1.34 in-sample. A bootstrap test returned p = 0.774 (not significant). Walk-forward Sharpe stayed positive in 2 of 5 folds. Worst out-of-sample drawdown was -30.49%.

Analysis

In-sample the fade loses badly: over 675 trading days the after-cost IS Sharpe is -1.34 with a -69.7% total return and a -73.8% drawdown. There is no promising signal to degrade from — the hypothesis is contradicted from the outset, which is the cleanest kind of rejection.

Out of sample the picture is the same. Across 407 independent intraday trades the per-trade after-cost Sharpe is -0.63 (win rate 45.0%), and the daily mark-to-market OOS Sharpe is -0.35 with a -21.5% total return and -30.5% drawdown. The strategy is systematically on the wrong side: large gaps continue (gap-and-go) far more than they revert, so fading them bleeds capital both in and out of sample.

Walk-forward confirms instability rather than edge: fold Sharpes are -0.67, -0.30, +0.81, -1.16, +0.71, an oos_consistency of just 40% (2 of 5 folds positive), well below the 60% bar. The two positive folds are not enough to overturn the strongly negative aggregate.

The block-bootstrap significance test returns p = 0.774 on the OOS per-trade returns — the observed negative Sharpe is entirely consistent with the zero-edge null, so there is no evidence of any exploitable structure (in either direction) once costs are applied. Costs matter but are not the story here; the raw directional call is wrong.

Verdict: REJECTED. Trade count is ample (407 OOS trades, well above the 30 minimum), so this is a genuine rejection rather than an inconclusive small-sample result. The documented positive overnight-drift effect does not imply mean-reversion of gaps intraday; if anything QQQ gaps trend. No promotion, no paper trade.

Provenance

Synthesized from 1unkju6, 1ui54pp, 1uk06py. The June-26 SPY tape thread and the general overnight-drift discussion both single out the open as a special moment; the flip side of positive overnight drift is that outsized overnight GAPS tend to partially mean-revert during regular trading hours. Fading the gap is a distinct, testable edge from simply holding overnight (already tracked).

Evidence

Walk-forward Sharpe by fold

Each fold retrains on past data then tests on the next unseen window. Staying positive across folds is the real test of an edge.

---

_Generated by labs-algo-trading. Automated research — not financial advice. Backtests overfit; treat verdicts as hypotheses._