Volatility-Gated Nasdaq Week Trade

_Momentum-continuation payoffs concentrate in higher-volatility regimes; in quiet weeks mean-reversion dominates and the directional bet decays. Gating the same Wed->Fri directional trade on above-median trailing vol should raise per-trade expectancy if the effect is real._

Hypothesis

The mid-week Nasdaq continuation trade (long/short into Friday depending on where price sits versus Monday's open at Wednesday's close) carries positive expectancy only in weeks where trailing 20-day realised volatility is above its trailing one-year median.

Method

Instrument: QQQ · Entry: At Wednesday close take direction = sign(Wed close - Monday open); enter only if trailing 20d realised vol > trailing 252d median vol · Exit: Friday close

Results

Out-of-sample Sharpe was -0.85, versus 0.35 in-sample. A bootstrap test returned p = 0.864 (not significant). Walk-forward Sharpe stayed positive in 3 of 5 folds. Worst out-of-sample drawdown was -34.26%.

Analysis

In-sample the volatility-gated variant looked mildly encouraging — an IS Sharpe of 0.35 on 656 mark-to-market days (22.3% total return) — which is exactly the trap this pipeline is built to catch. Out of sample the edge inverts completely: OOS Sharpe -0.855, total return -28.5%, maximum drawdown -34.3% over 144 independent weekly trades. The gate did not rescue the underlying signal; it simply fit the in-sample window.

The out-of-sample directional hit rate was 54.2%, marginally above a coin flip, yet the mean per-trade return after costs was -0.20%. As with the ungated parent idea, a slightly-better-than-even hit rate is more than offset by negative payoff asymmetry: the losing weeks (including an -11.3% short-vol blow-up in early April 2026) are larger than the winners, so directional accuracy never converts into profit.

Walk-forward validation was inconsistent and dominated by a single lucky fold: fold Sharpes were [-1.06, 7.58, 0.91, 1.78, -0.88] for 60% positive folds, but the 7.58 fold carries the average and the combined walk-forward Sharpe is a negligible 0.10. This is the fingerprint of noise, not a stable regime edge.

The block-bootstrap significance test on non-overlapping OOS trades gives p = 0.864 under the zero-edge null — the observed OOS Sharpe is well within what pure chance produces. Costs are modelled conservatively (10 bps charged once per trade, capital marked idle between trades), and even so the strategy loses money; a heavier microstructure cost assumption would only deepen the loss.

Verdict: REJECTED. The volatility filter added a degree of freedom that flattered the in-sample fit and provided no genuine out-of-sample protection. Combined with the outright rejection of the ungated parent, there is no credible mid-week Nasdaq continuation edge in this data — the original Reddit claim of a ~76% weekly hit rate does not survive honest, cost-aware, out-of-sample testing.

Provenance

Synthesized from 1uhywj4, 1tjgscr. Weekly directional-continuation edges are regime dependent; conditioning the mid-week momentum trade on elevated realised volatility (when trends are more likely to extend and drawdowns cluster) is a credible refinement rather than speculative filler.

Evidence

---

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