How the TQQQ / SQQQ Tactical Strategy Works (2010–Present)
This page runs the same low-turnover Flagship rules as the 1985–Present Synthetic Backtester, but executes them on real adjusted daily closes for TQQQ, SQQQ and TMF since February 2010, so ETF expense ratios, swap financing, tracking error and compounding are already in the prices. Every signal (trend, RSI, volatility) is computed from TQQQ's underlying index, the Nasdaq-100 (^NDX), with no look-ahead: a signal at the close of day t governs the return on day t+1.
The sliders start at the published defaults (150 / 30 / 25 / 25 / 50). Auto-Tune runs on load and whenever you change its goal or window, and moves the sliders to the best setting it finds for that goal. Change a slider by hand and the page tells you it is in custom mode until you re-tune or reset.
The Rules
- 150-day ^NDX SMA trend core (±3% hysteresis): Hold TQQQ while ^NDX is more than 3% above its 150-day SMA; flip to SQQQ only when it closes more than 3% below. The 6% deadband avoids constant flip-flopping around the average.
- Oversold RSI(10) bounce (< 30) with a 2% stop: Inside a bear regime, a 10-day Wilder RSI below 30 on ^NDX means capitulation selling, so the strategy temporarily holds TQQQ for the snap-back. It returns to the bull regime if ^NDX reclaims the SMA, and exits to SQQQ if the bounce fails and ^NDX drops 2% from entry.
- Volatility cap (25%) with a 25-point rebalance deadband: When 10-day annualized volatility exceeds the cap, equity exposure scales to
min(100%, cap / vol)(floored at 15%) and the rest sits in TMF. Weights change only when the target moves by 25+ points, which keeps turnover low. - Bubble trim (25%) and crisis shield (50%): TQQQ is capped at 50% (the rest in TMF) when ^NDX is more than 25% above its SMA, and the whole portfolio moves to TMF when 10-day volatility exceeds 50%.
- Crisis panic buy (RSI < 35 for 2 days): While the crisis shield has the portfolio in TMF, if 10-day RSI on the index stays below 35 for two days in a row (a capitulation low), the strategy moves half the portfolio back into the 3x long ETF (50% long / 50% TMF) until the shield ends. In backtests this adds a few points of CAGR, mostly from the V-shaped rebounds of 2000–02, 2008 and 2020, without deepening drawdowns; the edge is modest, in-sample, and depends on a handful of crisis episodes.
- Optional taxable account: Off by default (Roth / IRA). When on, the engine tracks FIFO tax lots and settles tax each year at historical US federal short- and long-term rates plus NIIT, with loss carryforward.
How Auto-Tune Picks Settings
- Search: It scores several thousand combinations of SMA length, RSI level, volatility cap, bubble trim and crisis shield, then applies the best one.
- Goals: Maximum Raw CAGR chases the highest return. Drawdown Shield looks for the shallowest drawdown among settings that still beat buy-and-hold.
- Windows: All-Weather blends the full history (40%) with the last 10 years (60%) and penalizes settings that lose to buy-and-hold recently. The other windows score only the full history, the last 10 years, or whatever date range the chart is zoomed to.
- It is in-sample: Auto-Tune fits the history it can see. The warning line under the summary flags any setting that fails to beat buy-and-hold over the last 10 years.
What the 2010–Present Data Says
We re-ran the rules one setting at a time on 2010–present ^NDX using a replica of this engine with modeled 3x returns (real-ETF results run a few points higher; for example buy-and-hold 3x is about 34% a year in the replica versus about 37% for actual TQQQ). The direction matters more than the exact numbers. Buy-and-hold 3x returned about 34% a year with a roughly −82% worst drawdown; the defaults returned about 62% with −55%.
- Every trend length beat buy-and-hold, but 150 is a peak, not a plateau: 100/125/175/200/220-day SMAs returned about 45–54% a year, while 150 returned about 62%. Expect real-world results nearer the 45–55% range than the headline number, and treat 150 as a good choice rather than a magic one.
- The RSI bounce is the biggest single lever: With it off, CAGR fell to about 11% and drawdown deepened to −80%. Threshold 30 returned about 62%, but 25 returned about 26% and 35–40 about 41–46%, so this setting is sensitive and worth leaving alone.
- Vol cap is a risk-versus-return dial: No cap or a 35% cap returned about 67–68% with a −62% drawdown; the 25% cap returned about 62% with a −55% drawdown; a 15% cap cost about 12 points of return for no drawdown benefit.
- Crisis shield works best at 40–50%: Turning it off cost about 3 points of return, and 35% or 65% were worse than 40–50%.
- Bubble trim barely matters: 15% to off moved CAGR by only about 1–2 points because the 2010s never produced the 1990s-style parabolic stretches that make it bite.
- What 2010–present cannot tell you: This window is one long bull market with a handful of sharp dips (2011, 2018, 2020, 2022). See the Synthetic page for 1987, the dot-com bust and 2008.
Longer-History Robustness Notes (1985–Present)
Earlier tests on 1985–present ^NDX used a simplified 3x simulation (TMF sleeve treated as cash, rough financing costs), changing one thing at a time. Absolute returns differ from the backtester above, so treat the direction of each result as the takeaway.
- RSI threshold: 28–31 was a stable plateau (about 43–48% CAGR). Below 27 it collapsed (17–34% at 24–27) because too few bounces fire; above 32 it faded to about 35–41%.
- 2% bounce stop: A local optimum. Tighter stops (0.5–1.5%) get chopped out by normal rebound noise; wider stops (3–10%) push max drawdown to −90% or worse.
- No “structural bear” filter: Switching the bounce off during structural bears (falling 200-day SMA, death cross, or more than 25% below the 1-year high) cut CAGR from about 48% to 20–47% and deepened drawdowns. Bear-market bounces win less often (about 27%), but the winners are large and the 2% stop caps the losers.
- Single 150-day SMA ±3%: Beat 100/125/175/200-day alternatives and 15 multi-SMA ensembles (majority vote, unanimous, averaged), which lag turning points.
- No VIX filter: Requiring a VIX spike or roll-over before taking the bounce cut CAGR to roughly 6–35%. RSI already captures the capitulation.
- Whipsaws in V-shaped dips: Over the full 1985–present history, shorting SQQQ only when ^NDX is also below its 200-day SMA (otherwise holding TMF) added about 3–5 points of CAGR, almost all of it before 2010, which is why it does not change the results on this page. Details are in the Synthetic page's explainer.