QQQ Two-Day Dip Reversion

Conclusion: Inconclusive

Out-of-sample Sharpe was 0.89, versus -0.04 in-sample. A bootstrap test returned p = 0.000 (statistically significant).

Despite clearing the trade-count, Sharpe and p-value gates on the OOS slice, the strategy loses money and drops 75% in-sample. It is a bull-market beta harvester dressed as mean reversion, and it is not promoted.

Method

Rule / pseudocode

DATA ← adjusted observations for the stated instrument and horizon

RULE ← Buying QQQ after two or more consecutive down days and exiting after a fixed 3-day hold captures short-term mean reversion in the Nasdaq-100 with positive out-of-sample returns.

METHOD ← Instrument: QQQ · Entry: 2+ consecutive down closes

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

_Short-horizon equity mean reversion is a classic stat-arb style edge; the 'quant bloodbath last week of June' thread suggests this edge is crowded and regime-sensitive, making an honest OOS test valuable to see whether it has decayed._

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

0.89

Ann. return

14%

Max drawdown

-21%

Win rate

19%

p-value

significant

WF folds +

5/5

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

Buying QQQ after two or more consecutive down days and exiting after a fixed 3-day hold captures short-term mean reversion in the Nasdaq-100 with positive out-of-sample returns.

Method

Instrument: QQQ · Entry: 2+ consecutive down closes

Results

Out-of-sample Sharpe was 0.89, versus -0.04 in-sample. A bootstrap test returned p = 0.000 (statistically significant). Walk-forward Sharpe stayed positive in 5 of 5 folds. Worst out-of-sample drawdown was -20.89%.

Analysis

The out-of-sample numbers look superb in isolation — a Sharpe of 0.89, a 194.7% return over 225 non-overlapping trades, a -20.9% drawdown and p = 0.00 — but they collapse the moment they are set against the in-sample period. In-sample the same rule earned a Sharpe of -0.04 on a total return of -12.4% with a catastrophic -75.5% drawdown.

That IS/OOS reversal is the entire story. The in-sample window contains the dot-com crash and the 2008 crisis, during which 'buy Nasdaq after two down days' meant catching falling knives all the way down; the out-of-sample window (2015-2026) is a near-uninterrupted dip-buying bull market in which the same reflex was richly rewarded. The edge is therefore sample-window dependent, exactly the failure mode the pipeline is built to catch.

Walk-forward showed rising fold Sharpes (0.10, 0.52, 0.55, 0.83, 1.41) — monotonically improving toward the present, which is not stability but a time trend: the strategy has been getting 'better' only because markets have become more dip-buy-friendly, with no guarantee that persists.

The block-bootstrap p-value is significant, but significance here only confirms the OOS mean return is non-zero in a favourable regime; it says nothing about robustness across regimes, and the CLAUDE.md rule is explicit that an edge whose sign flips with the IS/OOS split is INCONCLUSIVE regardless of how good the OOS Sharpe looks.

Verdict: INCONCLUSIVE. Despite clearing the trade-count, Sharpe and p-value gates on the OOS slice, the strategy loses money and drops 75% in-sample. It is a bull-market beta harvester dressed as mean reversion, and it is not promoted.

Source

Originated from discussion on r/manual.

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._