SignalQuarryGitHub
← Strategy library

Leveraged ETF research · Research strategy

L04 · Covariance-controlled leveraged mix

Sets broad and semiconductor weights from trend eligibility and estimated joint volatility.

TQQQ or SPXL, SOXL, and settled cash; QQQ or SPY and SOXX are signal proxiesWeeklyLeveraged ETF exposure

The idea

How it works

Sets broad and semiconductor weights from trend eligibility and estimated joint volatility.

Why it might work

Accounting for correlation may size a two-fund portfolio more cautiously.

When it invests

At weekly reviews, start near 60% broad and 40% semiconductor, then scale both toward a 40% portfolio volatility target.

When it exits

Exit any fund whose proxy falls below its 200-day average.

What can go wrong

  • Historical signals can reverse before a delayed order fills.
  • Leveraged fund losses can be large and fast.

Selected research result

L04 · TQQQ + SOXL · v3 base backtest

L04 on TQQQ + SOXL passed the repaired v3 study's historical qualification gates. This is a simulated research result, not live or paper authorization.

Total return233.9%Change in simulated portfolio value over the test.
Annualized return49.5%Yearly compounded return; shown only when supported.
Worst drawdown33.7%Largest fall from an earlier portfolio high.
Sharpe ratio1.22Return relative to variability, under the source study's definition.
Trades9Completed simulated trades, when reported.

Read assumptions and limitations →

Research history

Variants and recorded runs

Repeated and diagnostic runs remain accessible under their tested variant. No run is selected by highest return.

SPXL SOXL1 result
TQQQ SOXL1 result