TQQQ SMA Trend Strategy vs Jason Kelly’s 9Sig: why I think this may be a more robust approach
I have been researching systematic ways to hold leveraged tech exposure, especially TQQQ, while reducing the risk of catastrophic drawdowns.
A popular approach is Jason Kelly’s 9Sig strategy, which uses TQQQ with a value-averaging / rebalancing framework. The general idea is to buy more TQQQ after declines and trim after strong gains. I understand the appeal: TQQQ is extremely volatile, and systematically buying weakness can work very well if the asset eventually recovers.
However, the main issue I have with that approach is that it remains structurally dependent on TQQQ recovering. In a long Nasdaq bear market, or a regime similar to 2000–2013, continuing to buy TQQQ weakness could be psychologically and financially brutal.
So I have been testing a different framework:
Strategy
- When TQQQ is above its 150-day simple moving average, hold 100% TQQQ.
- When TQQQ is below its 150-day simple moving average, rotate out of TQQQ and hold:
- 30% KMLM
- 30% DBMF
- 40% UGL
The idea is simple: when the Nasdaq/TQQQ trend is intact, stay fully aggressive. When the trend breaks, do not average down into TQQQ; move into assets that can perform for different macro reasons.
Why these defensive assets?
KMLM and DBMF are managed-futures / trend-following ETFs. They can potentially benefit from trends in commodities, currencies, bonds, and rates. They are not dependent on Nasdaq earnings or tech multiples.
UGL is 2x gold exposure. It is not a perfect hedge, but it can help in inflationary, monetary-stress, dollar-weakness, or crisis regimes.
So the “risk-off” basket is not cash only, and it is not another hidden tech bet. It is designed to move away from tech beta when tech momentum breaks.
Why I prefer this over a 9Sig-style TQQQ strategy
Jason Kelly’s 9Sig logic is essentially to exploit TQQQ volatility by buying lower lows and selling higher highs. That can work well in a volatile bull market or in sharp drawdowns followed by strong recoveries.
But my concern is what happens in a long bear market.
With 9Sig, the strategy can keep adding exposure to a falling leveraged ETF. That may be fine if the bear market is short. But if the bear market lasts years, the investor needs enormous discipline and enough dry powder to keep buying.
My strategy does the opposite. It says:
>
That seems conceptually more robust to me.
Backtest result
Using the available data from late 2020 to July 2026, the SMA150 version performed better than the same system using SMA200.
Approximate result for the strategy:
- TQQQ > SMA150: 100% TQQQ
- TQQQ < SMA150: 30% KMLM / 30% DBMF / 40% UGL
In my backtest, $100,000 grew to roughly $708k, with a CAGR around 42%, a max drawdown around -33%, and a Calmar ratio around 1.28.
The SMA200 version was still good, but weaker: lower CAGR and higher drawdown.
I also tested replacing TQQQ with SOXL. SOXL produced much higher returns in the recent sample, but I consider it conceptually weaker as a core strategy because it is much more concentrated in semiconductors, more cyclical, more volatile, and more sensitive to the exact moving-average parameter.
My conclusion
I am not saying this is perfect or future-proof. The data window is short because KMLM and DBMF do not have very long histories. Also, this strategy still uses TQQQ, so it is obviously very aggressive.
But conceptually, I prefer this framework to a pure 9Sig-style TQQQ averaging strategy because it separates regimes:
- In a strong tech bull market: hold TQQQ.
- In a broken tech trend: leave TQQQ and rotate to macro/trend/gold exposure.
- Re-enter TQQQ only when the trend recovers.
To me, that seems more robust than continuing to buy more TQQQ during a potentially long bear market.
Curious to hear criticism, especially from people who have studied 9Sig, trend-following, or leveraged ETF decay.
