13 Strategies Tested: TQQQ 583%, QLD 336%
"If it's not optimal, don't step into the market."
In Part 2, we proved that a 200-day MA + stop-loss strategy decisively outperforms simple buy-and-hold. Just a 200-day moving average condition and a holding period limit produced a 35x difference in returns for a 3x ETF.
But the Part 2 strategy had its limitations. First, fixed holding periods (6 weeks for 2x, 2 weeks for 3x) force you to liquidate even when a strong uptrend continues. You can't ride the trend to its end. Second, the 200-day MA is a long-term indicator that reacts slowly at the onset of a decline. Third, the stop-loss thresholds (-20%/-22%) are wide enough that you only exit after suffering significant losses.
This installment overcomes those limitations. Instead of fixed holding periods, we switch to moving average-based trend following -- holding as long as the trend is alive. We use shorter moving averages (10-week, 5-week) to respond quickly to declines, and tighten stop-losses to -15%/-10% to preserve capital. Three strategies for 1x QQQ, five for 2x QLD, and five for 3x TQQQ -- a total of 13 strategies exhaustively tested across 5.8 years of weekly data to find the optimal combination considering total return, Sharpe ratio, and maximum drawdown.
Let me give you the conclusion upfront: the optimal strategy is completely different depending on the leverage multiplier.
Test Setup
All strategies were tested under identical conditions.
Backtesting Environment
- Period
- Feb 2020 - Oct 2025 (approx. 5.8 years, 300 weeks)
- Target ETFs
- QQQ (1x), QLD (2x), TQQQ (3x)
- Strategies Tested
- 13 total (1x: 3, 2x: 5, 3x: 5)
- Initial Capital
- $100,000 per strategy
- Data Basis
- Weekly closing prices
- Evaluation Metrics
- Return, Sharpe ratio, MDD, win rate
The key evaluation metric is the Sharpe ratio. No matter how high the returns, if volatility is too great, you won't survive in practice. How much return you earn per 1% of risk -- that's the true measure of a strategy's merit.
1x QQQ: Do Nothing at All
We tested three strategies for QQQ at 1x leverage.
| Strategy | Return | Sharpe Ratio | MDD | Trades | Win Rate |
|---|---|---|---|---|---|
| E1: Buy & Hold | 168.10% | 0.723 | -35.46% | 1 | - |
| E2: 50-Week MA | 94.43% | 0.759 | -20.10% | 9 | 100% |
| E3: Golden Cross 10/20 | 58.36% | 0.523 | -15.19% | 9 | 50% |
The results are unambiguous. Buy-and-hold dominates with a 168% return.
The 50-week MA strategy has a lower MDD of -20% and a slightly higher Sharpe ratio of 0.759. But the return stops at 94%. Is a 15 percentage-point improvement in MDD worth giving up 73 percentage points of return?
The golden cross strategy has a win rate of just 50% and the lowest return at 58%. Using technical indicators on a 1x ETF actually hurts performance.
Why Buy-and-Hold for 1x
A 1x ETF has no volatility decay. If QQQ drops 10% and then bounces 11.1%, you're back to even. This simple math is what makes long-term holding viable.
Moving averages can help you time entries and exits, but they simultaneously create re-entry failure risk. You sell when the price drops below the 50-week MA, but if it sharply rebounds, you're forced to buy back at a higher price. This opportunity cost created the 73 percentage-point return gap over 5.8 years.
1x QQQ Conclusion
- Optimal Strategy
- E1: Buy & Hold
- Return
- 168.10% (18.64% annualized)
- Core Philosophy
- "Time is your friend" - doing nothing is the best move
- Execution Difficulty
- Minimal - buy and forget
2x QLD: The 10-Week MA Is the Golden Threshold
We tested five strategies for QLD at 2x leverage. The results are dramatic.
QLD 2x Leverage Strategy Comparison
The 10-week MA + 15% stop-loss strategy (F2) dominates across every metric. Its 336% return is nearly double the runner-up (189%), and its Sharpe ratio of 0.977 is a phenomenal figure approaching 1.0.
Detailed Performance Comparison
| Strategy | Return | Annualized | Sharpe Ratio | MDD | Trades | Win Rate |
|---|---|---|---|---|---|---|
| F2: 10-Wk MA + 15% Stop-Loss | 336.31% | 29.09% | 0.977 | -30.85% | 43 | 52.4% |
| F5: 10+20 Wk Dual | 189.01% | 20.20% | 0.716 | -30.65% | 39 | 52.6% |
| F1: 20-Wk MA + 10% Stop-Loss | 177.62% | 19.36% | 0.626 | -34.79% | 35 | 58.8% |
| F4: 20-Wk MA + 8-Wk Limit | 129.45% | 15.48% | 0.510 | -29.34% | 55 | 51.9% |
| F3: Golden Cross 5/20 + Stop-Loss | 98.22% | 12.59% | 0.438 | -45.29% | 15 | 57.1% |
Why the 10-Week MA
The moving average period is a trade-off between sensitivity and stability.
| MA Period | Characteristics | Suitability for 2x Leverage |
|---|---|---|
| 5-Week | Too sensitive, excessive false signals | Inefficient due to overtrading |
| 10-Week | Adequate signal filtering, captures medium-term trends | Optimal |
| 20-Week | Slow signals, missed opportunities | Late entries and exits |
The 10-week MA reflects roughly 2.5 months of trend data. Given the volatility profile of 2x leverage, this period hits the sweet spot. Go too short and you get whipsawed by noise; go too long and you respond too late to declines, allowing volatility decay to compound.
In fact, the 20-week MA strategy (F1) returned only 177%. Even with the same stop-loss (10%), late entries and exits caused missed profit opportunities.
Why 15% Stop-Loss
The stop-loss threshold must be calibrated to the leverage multiplier.
If QQQ drops -7.5% -> QLD drops roughly -15%
A -7.5% move in QQQ is a "correction" -- not a trend reversal. Therefore, a 15% stop-loss on QLD weathers normal corrections while avoiding real bear markets.
A 10% stop-loss is too tight. It triggers on routine QLD fluctuations, causing a surge in unnecessary trades. A 20% stop-loss is too loose. A -20% drawdown on 2x leverage requires a +25% recovery to break even -- a substantial burden.
15% is the sweet spot between the two.
What a 0.977 Sharpe Ratio Means
Sharpe Ratio = (Return - Risk-Free Rate) / Volatility = (29.09% - 2%) / 27.72% = 0.977
Interpretation: 0.977% excess return per 1% of volatility (risk)
The general interpretation of the Sharpe ratio is as follows:
| Sharpe Ratio | Interpretation | Example |
|---|---|---|
| 0.0 - 0.3 | Questionable investment value | Buy-and-hold 3x ETF |
| 0.3 - 0.5 | Below average | Most individual stocks |
| 0.5 - 0.7 | Good | Well-diversified portfolio |
| 0.7 - 1.0 | Excellent | QLD 10-week MA strategy (0.977) |
| 1.0+ | Outstanding | Top-tier hedge funds |
The QLD 10-week MA strategy's 0.977 means you earn nearly 1% of excess return for every 1% of risk you take. This is a level most professional fund managers struggle to achieve.
2x QLD Conclusion
- Optimal Strategy
- F2: 10-Week MA + 15% Stop-Loss
- Return
- 336.31% (29.09% annualized)
- Sharpe Ratio
- 0.977 (near-perfect risk-adjusted return)
- Core Philosophy
- "The trend is your friend" - hold as long as the trend is alive
QLD 2x - 10-Week MA + 15% Stop-Loss Strategy (Weekly, 2020-2025)
Weekly candlestick chart. Green arrows = buy, red arrows = sell. Blue line = 10-week moving average. Buy on breakout above the 10-week MA; liquidate on breakdown or 15% drawdown.
3x TQQQ: Speed Is Survival
We also tested five strategies for TQQQ at 3x leverage. Here, the pattern shifts dramatically.
TQQQ 3x Leverage Strategy Comparison
The 5-week MA + 10% stop-loss strategy (G1) dominates with a 583% return. Over 5.8 years, $100,000 grows to $683,400.
Detailed Performance Comparison
| Strategy | Return | Annualized | Sharpe Ratio | MDD | Trades | Win Rate |
|---|---|---|---|---|---|---|
| G1: 5-Wk MA + 10% Stop-Loss | 583.44% | 39.53% | 0.895 | -51.31% | 67 | 54.5% |
| G4: 5-Wk MA + 4-Wk Limit | 451.40% | 34.44% | 0.825 | -56.11% | 99 | 57.1% |
| G5: 10-Wk MA + 6-Wk Limit | 243.82% | 23.87% | 0.567 | -50.52% | 75 | 56.8% |
| G3: Golden Cross 5/10 + Stop-Loss | 164.95% | 18.40% | 0.406 | -68.07% | 29 | 42.9% |
| G2: 2-Wk Momentum | 131.68% | 15.68% | 0.471 | -40.16% | 77 | 52.6% |
Why the 5-Week MA
3x leverage is an entirely different world from 2x.
QQQ -5% -> TQQQ approx. -15% QQQ -10% -> TQQQ approx. -30% QQQ -20% -> TQQQ approx. -60%
A 10% decline in QQQ can happen in just a few weeks. In that time, TQQQ evaporates by 30%. A slow indicator like the 10-week MA (2.5 months) only generates a sell signal after a significant portion of your capital has already vanished. In fact, the 10-week MA + 6-week limit strategy (G5) returned only 244% -- less than half of the 5-week MA strategy.
The 5-week MA reflects roughly 1.2 months of trend data. It reacts quickly while still filtering out week-to-week noise.
The Meaning of a Strict 10% Stop-Loss
For a 3x ETF, a 10% stop-loss is not a luxury -- it's a survival strategy. The loss recovery formula from Part 2 proves why.
| Drawdown | Return Required to Recover |
|---|---|
| -10% | +11.1% |
| -20% | +25.0% |
| -30% | +42.9% |
| -50% | +100.0% |
If TQQQ drops -30%, you need +43% just to get back to breakeven. Given 3x leverage volatility, that's extremely difficult. But if you cut at -10%, you only need +11% to recover.
In the G1 strategy, the 10% stop-loss triggered on roughly 18 out of 67 trades (27%). Those 18 stop-losses preserved capital, creating the opportunity for the remaining 49 trades.
3x TQQQ Conclusion
- Optimal Strategy
- G1: 5-Week MA + 10% Stop-Loss
- Return
- 583.44% (39.53% annualized)
- Core Philosophy
- "The stop-loss is your friend" - fast exits create re-entry opportunities
- Caution
- MDD of -51.31% poses significant psychological strain
TQQQ 3x - 5-Week MA + 10% Stop-Loss Strategy (Weekly, 2020-2025)
Weekly candlestick chart. Green arrows = buy, red arrows = sell. Blue line = 5-week moving average. Buy on breakout above the 5-week MA; liquidate immediately on breakdown or 10% drawdown.
Key Finding: Strategy Philosophy by Leverage Level
One pattern runs through all 13 strategies: the higher the leverage, the faster everything must be.
| Leverage | Optimal MA | Optimal Stop-Loss | Avg. Holding Period | Philosophy |
|---|---|---|---|---|
| 1x (QQQ) | None (Buy & Hold) | None | Unlimited | "Time is your friend" |
| 2x (QLD) | 10-Week (2.5 months) | 15% | ~4 weeks | "The trend is your friend" |
| 3x (TQQQ) | 5-Week (1.2 months) | 10% | ~2.6 weeks | "The stop-loss is your friend" |
The root cause of this pattern is volatility decay.
Annual Volatility Decay = L(L-1) x sigma^2 / 2
1x (L=1): 1(1-1) x sigma^2 / 2 = 0 (no volatility decay) 2x (L=2): 2(2-1) x sigma^2 / 2 = sigma^2 3x (L=3): 3(3-1) x sigma^2 / 2 = 3 x sigma^2
The volatility decay of a 3x ETF is 3 times that of a 2x ETF.
A 1x ETF has zero volatility decay, so compounding works purely in your favor over time. This is the mathematical reason buy-and-hold is optimal.
A 2x ETF experiences volatility decay, but at a manageable level. By using the 10-week MA to avoid major declines, trend momentum offsets the volatility drag.
A 3x ETF suffers the most punishing volatility decay. You must react quickly with the 5-week MA and defend your capital with a 10% stop-loss. Even a slight delay leads to irrecoverable losses.
Return vs. Risk: What Will You Choose?
Here is a comprehensive comparison of risk and return across the three optimal strategies.
Optimal Strategy Comparison by Leverage Level
There is a remarkable finding here.
The 2x QLD strategy delivers double the return of 1x (168% vs. 336%), yet its MDD is actually lower (-35% vs. -31%). In terms of risk-adjusted efficiency, the QLD 10-week MA strategy is the clear winner. The 0.977 Sharpe ratio proves it.
On the other hand, 3x TQQQ posts a staggering 583% return, but with a -51% MDD. That means $100,000 drops to $49,000 at some point. Very few investors can withstand that kind of pain.
Selection Guide by Investor Profile
- Stability-Oriented
- QQQ 1x Buy & Hold - 18.64% annualized, no maintenance required
- Efficiency-Oriented
- QLD 2x 10-Week MA - 29.09% annualized, check once a week (Recommended)
- Return-Maximizing
- TQQQ 3x 5-Week MA - 39.53% annualized, check once a week + strict stop-losses
Portfolio Combination Simulation
Instead of a single strategy, combining them can diversify risk. Here are the projected outcomes over 5.8 years on a $100,000 initial investment.
| Portfolio | Allocation | Est. Return | Est. MDD | Maintenance Level |
|---|---|---|---|---|
| Conservative | QQQ 70% + QLD 30% | ~219% | ~-34% | Low |
| Balanced | QQQ 40% + QLD 40% + TQQQ 20% | ~318% | ~-38% | Moderate |
| Aggressive | QLD 50% + TQQQ 50% | ~460% | ~-41% | High |
The balanced portfolio can be expected to deliver approximately 26.6% annualized returns, with MDD contained to around -38%. In particular, the 40% allocation to QLD serves as the key driver lifting the overall portfolio efficiency.
Limitations of Backtesting
These results should not be taken as gospel.
Important Caveats
- Period Bias
- 2020-2025 is centered on the post-COVID recovery and bull market
- Transaction Costs
- Commissions, taxes, and slippage not reflected (real returns would be 10-15% lower)
- Psychological Factors
- Backtesting assumes mechanical execution; real trading involves emotions
- Future Uncertainty
- Past patterns are not guaranteed to persist in the future
In particular, the 3x TQQQ strategy is predicated on mechanically enduring a -51% MDD. Very few investors can stay convinced that "it'll come back" the moment $100,000 has shrunk to $49,000. A strategy's theoretical superiority and its practical executability are two entirely different matters.
Conclusion: The Higher the Multiplier, the Sharper the Knife's Edge
Leveraged ETFs are like knives. 1x is a butter knife, 2x is a chef's knife, 3x is a sashimi blade. The sharper the tool, the more precise your technique must be.
With 1x, just hold on. With 2x, you must read the trend. With 3x, you need to learn how to escape before anything else.
In the next installment (Part 4), we'll cover how to execute these optimal strategies in practice with a concrete trading guide. From calculating moving averages, to weekly checklists, to lessons learned from 64 actual trades -- the final step in turning theory into practice.