Leverage Rotation Strategy (LRS): UPRO above the 200-day average

Jun 2009 – Sep 2026	Strategy	SPY
Annual return (CAGR)	24.6%	15.1%
Worst drawdown	-51.9%	-33.7%
Sharpe ratio	0.79	0.91
Volatility	36.1%	17.0%
Annual return since publication (Mar 2016)	26.9%	15.6%

Jan 1960 – Sep 2026	Strategy	SPY
Annual return (CAGR)	16.6%	10.6%
Worst drawdown	-72.0%	-55.2%
Sharpe ratio	0.63	0.70
Volatility	32.9%	16.4%
Annual return since publication (Mar 2016)	26.9%	15.6%

The Leverage Rotation Strategy (LRS) comes from Michael Gayed and Charles Bilello's paper "Leverage for the Long Run", which won the 2016 Charles H. Dow Award. It uses leverage only while the stock market is trending up: hold a leveraged S&P 500 fund while the S&P 500 closes above its 200-day moving average, and Treasury bills while it closes below. This page covers the three-times version with UPRO; the two-times version uses SSO.

The idea#

A fund that resets its leverage every day does not decay by itself, the authors argue. What hurts it is volatility, and especially markets that alternate between up and down days. A steady run of gains compounds faster than three times the index; a choppy market grinds the fund down even when the index goes nowhere.

Gayed and Bilello found that the S&P 500 behaves differently on either side of its 200-day average. Above it, the market has tended to be calmer and to rise in longer streaks, the conditions in which daily leverage works. Below it, volatility rises and streaks break down. So the strategy takes leverage only above the average and waits in Treasury bills below it.

How it works#

At the close of every trading day:

Compare the S&P 500 (SPY) with its 200-day simple moving average.
If SPY closes above its average, hold UPRO, a fund that aims to return three times the S&P 500's daily move.
Otherwise, hold Treasury bills (BIL).

The paper also tested shorter averages and 1.25x and 2x leverage. It focused on the 200-day average because it trades least often.

What the backtest shows#

UPRO launched in mid-2009, so the test with real prices starts then. Since then the strategy compounded near 25% a year, against about 15% for the S&P 500. It did so with twice the S&P 500's volatility, and on risk-adjusted return it trailed both the S&P 500 and a 60/40 portfolio. Its worst drawdown was about 52%, against about 34% for the S&P 500.

The results swing widely from year to year. It more than doubled in 2013 and nearly doubled in 2021, both calm, steadily rising years. It lost about 30% in 2011, when the S&P 500 gained 2%, and about 40% in 2022, more than twice the market's loss. The worst drawdown came quickly, from April to September 2010, when the market crossed its average back and forth and each wrong signal cost three times over. It took until May 2013 to recover.

Since the paper appeared in early 2016 the strategy has returned about 27% a year, but it still fell by about half at its worst point.

With simulated history the test starts in 1960. The model treats UPRO as three times the S&P 500's daily return, minus the fund's fee and the cost of borrowing at a little above the Treasury bill rate. Over that longer span it returned about 16.5% a year against about 10.5% for the S&P 500, but with a lower risk-adjusted return. It gained in 1974, when the market fell by a quarter, and rose 40% in 1987, when the market gained 5%. It lost 53% in 2000, as rallies pulled it back into a falling market. Its worst simulated drawdown was 72%, from mid-1999 to April 2003, and it did not regain that peak until late 2013.

When it struggles#

Choppy markets. When the S&P 500 hovers around its 200-day average, every false signal is taken with triple exposure, as in 2010, 2011, 2015, and 2022.
Crashes from a high. A sudden fall that starts well above the average hits the fund at full leverage before the signal turns.
Long bear markets with sharp rallies. Each rally above the average buys back in, and the next leg down takes three times its toll.

Using it on Tactfolio#

The live strategy runs these rules each day on SPY, UPRO, and BIL, with the fund's real fees and financing costs in place of the paper's flat leverage cost. Copy it to try SSO or SPXL, a different average, or a band around the average to reduce whipsaws.

Year by year

Year	Strategy	SPY
2026*	18.1%	14.0%
2025	23.5%	17.7%
2024	63.6%	24.9%
2023	30.6%	26.2%
2022	-39.7%	-18.2%
2021	98.6%	28.7%
2020	18.3%	18.3%
2019	44.4%	31.2%
2018	-12.6%	-4.6%
2017	71.4%	21.7%
2016	29.3%	12.0%
2015	-17.6%	1.2%
2014	31.7%	13.5%
2013	118.5%	32.3%
2012	37.1%	16.0%
2011	-29.7%	1.9%
2010	-9.3%	15.1%
2009*	73.8%	22.3%

Year	Strategy	SPY
2026*	18.1%	14.0%
2025	23.5%	17.7%
2024	63.6%	24.9%
2023	30.6%	26.2%
2022	-39.7%	-18.2%
2021	98.6%	28.7%
2020	18.3%	18.3%
2019	44.4%	31.2%
2018	-12.6%	-4.6%
2017	71.4%	21.7%
2016	29.3%	12.0%
2015	-17.6%	1.2%
2014	31.7%	13.5%
2013	118.5%	32.3%
2012	37.1%	16.0%
2011	-29.7%	1.9%
2010	-9.3%	15.1%
2009	72.5%	26.4%
2008	-10.5%	-36.8%
2007	-28.0%	5.1%
2006	15.3%	15.8%
2005	-22.2%	4.8%
2004	5.3%	10.7%
2003	62.3%	28.2%
2002	-18.2%	-21.6%
2001	3.3%	-11.8%
2000	-52.6%	-9.7%
1999	14.8%	20.4%
1998	36.0%	28.7%
1997	82.9%	33.5%
1996	42.0%	22.5%
1995	122.1%	38.0%
1994	-25.8%	0.4%
1993	18.3%	9.7%
1992	8.2%	7.6%
1991	63.2%	30.3%
1990	-35.2%	-3.2%
1989	72.2%	31.5%
1988	6.0%	16.4%
1987	39.8%	5.1%
1986	31.4%	18.6%
1985	81.1%	31.6%
1984	-4.1%	6.2%
1983	39.4%	22.4%
1982	68.0%	21.6%
1981	-35.6%	-4.9%
1980	51.7%	32.4%
1979	12.4%	18.3%
1978	-18.5%	6.5%
1977	-24.4%	-7.2%
1976	45.0%	23.8%
1975	49.0%	37.1%
1974	8.0%	-26.5%
1973	-35.7%	-14.7%
1972	46.5%	18.9%
1971	24.5%	14.1%
1970	38.1%	3.9%
1969	-22.7%	-8.4%
1968	18.3%	10.9%
1967	52.7%	23.8%
1966	-18.5%	-10.1%
1965	12.1%	12.4%
1964	42.0%	16.4%
1963	65.3%	22.7%
1962	-6.5%	-8.8%
1961	87.9%	26.8%
1960*	-5.8%	0.4%

* Partial year.

The rules as implemented

This is the exact tree Tactfolio runs, rebalanced daily with signals and trades at the close. Open it to inspect or copy it.

Sources and caveats

UPRO, a 3x daily-reset S&P 500 fund, stands in for the paper's 3x daily-leveraged S&P 500 total return with a 1% annual leverage cost; the fund's own fees and financing costs apply instead.
SPY stands in for the S&P 500 total return index that the 200-day average is taken on, and BIL for 3-month Treasury bills.
Signals and trades use each day's close.

Common questions#

What is the Leverage Rotation Strategy?#

It is a rule from Michael Gayed and Charles Bilello's 2016 paper "Leverage for the Long Run". You hold a daily leveraged S&P 500 fund while the S&P 500 is above its 200-day moving average and Treasury bills while it is below, so leverage is used only in calmer, rising markets.

Can you use the 200-day moving average with TQQQ instead of UPRO?#

The paper tested only the S&P 500, and this page does too. Traders often apply the same rule to the Nasdaq 100 with TQQQ; you can copy the strategy, replace SPY with QQQ and UPRO with TQQQ, and backtest that variant yourself.

How much can the Leverage Rotation Strategy lose?#

A lot. With real prices its worst drawdown was about 52%, in 2010, and in simulated history since 1960 it was 72%, from 1999 to 2003. The 200-day rule limits long bear markets but not fast falls or repeated whipsaws.

How is it different from the 200-day moving average strategy?#

The signal is the same. The 200-day moving average strategy holds the S&P 500 itself above the average, while the Leverage Rotation Strategy holds a two- or three-times leveraged fund, which multiplies both the gains and the whipsaw losses.