QQQ Donchian Breakout

Conclusion: Inconclusive

Out-of-sample Sharpe was 0.92, versus 0.24 in-sample. A bootstrap test returned p = 0.000 (statistically significant).

The idea reframes an opaque 'Fibo H4' pattern claim into a transparent, testable Donchian rule — which was worthwhile — but the honest reading is too few OOS trades, a weak in-sample Sharpe, and an OOS result that leans on a single favourable trend regime. It is not promoted; it would need more instruments or a longer OOS window before it could be judged.

Method

Rule / pseudocode

DATA ← adjusted observations for the stated instrument and horizon

RULE ← A Donchian-channel breakout (enter long when price closes above its trailing 50-day high, exit on close below its trailing 20-day low) on QQQ captures trend continuation with positive out-of-sample risk-adjusted returns.

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

_Reframes the poster's suspiciously-clean Fibo H4 pattern claim into a transparent, well-known breakout rule (Donchian/turtle) that can be honestly tested for overfitting, rather than trusting an opaque discretionary pattern._

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

0.92

Ann. return

11%

Max drawdown

-19%

Win rate

30%

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

A Donchian-channel breakout (enter long when price closes above its trailing 50-day high, exit on close below its trailing 20-day low) on QQQ captures trend continuation with positive out-of-sample risk-adjusted returns.

Method

Instrument: QQQ

Results

Out-of-sample Sharpe was 0.92, versus 0.24 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 -19.16%.

Analysis

The out-of-sample slice again flatters the rule — a Sharpe of 0.92 on a 139.4% return with a -19.2% drawdown — but it rests on only 21 completed round-trip trades. The pipeline's own promotion rule states that fewer than 30 independent OOS trades cannot be assessed no matter how good the Sharpe looks, so this alone caps the verdict at INCONCLUSIVE.

The in-sample evidence is also weak: a Sharpe of just 0.24 with a -49.4% drawdown across 57 trades. Breakout systems are known to whipsaw badly in range-bound and mean-reverting markets, and the deep in-sample drawdown is consistent with that. The OOS improvement to 0.92 is driven by the strong trending Nasdaq of 2018-2026 rather than by any change in the rule.

Walk-forward folds were all positive (0.22, 0.63, 0.51, 1.54, 0.93) for 100% consistency, which is mildly encouraging, but with so few trades per fold the fold Sharpes are estimated on a handful of round-trips each and are not statistically reliable.

The block-bootstrap returned a nominally significant p-value, but with 21 observations the test has little power and the estimate is fragile; it should not be over-weighted against the trade-count and in-sample evidence.

Verdict: INCONCLUSIVE. The idea reframes an opaque 'Fibo H4' pattern claim into a transparent, testable Donchian rule — which was worthwhile — but the honest reading is too few OOS trades, a weak in-sample Sharpe, and an OOS result that leans on a single favourable trend regime. It is not promoted; it would need more instruments or a longer OOS window before it could be judged.

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._