QQQ Golden Cross Backtest: 36% Max Drawdown vs 83% for Buy and Hold, No Extra Return
This strategy does not make more money than QQQ. What it does is cut the worst QQQ crashes by more than half. QQQ Golden Cross watches one thing: whether the 50-day moving average of QQQ sits above the 200-day. Above, it holds QQQ. Below, it sits in cash. Backtested from 1999 to 2026 at 0.1% per side for fees and slippage, its worst drawdown was 36.2%, against 83.0% for holding QQQ. The price is return. It made 10.4% a year against 10.3% for QQQ, and it trailed QQQ in most years after 2003. What a golden cross strategy does The 50-day average tracks the last two-plus months of prices. The 200-day covers roughly a year. When the short one moves above the long one, the medium-term trend has turned up. Traders call that a golden cross, and the reverse a death cross. The strategy uses the textbook 50 and 200 with no tuning for the backtest. We scanned the short average from 30 to 70 days and the long one from 150 to 250. Sharpe stayed between 0.51 and 0.64. The 50/200 setting happens to be the top cell, but the others are close behind. It checks once a day after the close. It only buys, never shorts and never uses leverage. Over 27 years it entered 17 times and was invested about 74% of the time. Every year of the backtest YearStrategyQQQ 1999 (from Mar 10)+1.6%+78.9% 2000+2.4%−36.1% 2001+0.0%−33.3% 2002+0.0%−37.4% 2003+44.5%+49.7% 2004+0.4%+10.5% 2005−8.0%+1.6% 2006−2.5%+7.1% 2007+18.9%+19.0% 2008−17.2%−41.7% 2009+26.8%+54.7% 2010+10.5%+20.1% 2011−8.1%+3.5% 2012+20.2%+18.1% 2013+31.6%+36.6% 2014+19.2%+19.2% 2015+3.0%+9.4% 2016−3.3%+7.1% 2017+32.6%+32.7% 2018+4.9%−0.1% 2019+17.2%+39.0% 2020+35.3%+48.4% 2021+27.3%+27.4% 2022−11.9%−32.6% 2023+43.7%+54.9% 2024+25.6%+25.6% 2025+3.2%+20.8% 2026 (to Nov 19)+5.2%+5.4% In 1999 it spent most of the year waiting for the averages to warm up and missed that rally. Backtest results do not promise future returns. Where it falls less In the 2000 to 2002 dot-com crash, QQQ fell 83% from its peak. The strategy lost 36% by May 2000, then the death cross took it out. It stayed in cash for more than two years and was back at its old high in 2003. That was its deepest drawdown, and it is the reason the strategy exists. In the 2008 crisis QQQ fell 53% at worst while the strategy fell 24%. In 2022, the rate-hike year, QQQ lost 32.6% and the strategy lost 11.9%. It does not dodge every fall. In the March 2020 crash it dropped 29%, about as much as QQQ. Prices fell too fast for the averages to cross, and by the time it exited the rebound had started. Where it loses It loses on return. A moving-average cross always lags price, so it exits after part of the fall and gets back in after part of the rise. In most years it made a little less than holding QQQ. It trailed by 22 points in 2019 and by 18 points in 2025. Sideways markets hurt too. When the averages keep crossing back and forth, it takes several small losing trades, as in 2005, 2011 and 2016. Costs add up as well. At 0.1% per side, fees took about 3.3 percentage points in total over 27 years. How live trading differs from the backtest The signal comes from QQQ daily bars, and the backtest fills at the QQQ price. Live, it trades the QQQ stock perpetual on an exchange. The perpetual trades around the clock at its own price and charges funding, which the backtest leaves out. Backtest prices are adjusted for splits and dividends, so the return includes dividends. The perpetual pays none. Who it suits People who want to own tech stocks but could not sit through a halving. It will not beat the index on return. It trades that for falling less than half as far in the worst case. If extra return is what you want, it does not deliver. The backtest made 10.4% a year against 10.3% for QQQ and trailed in most years. Running it yourself Add it to Blave Agent to rerun the backtest, change the averages, or try another fund such as SPY. It is a single-asset strategy, so it can trade the QQQ perpetual automatically on an exchange that offers it where you are. Read the live-trading differences above before you turn it on. See QQQ Golden Cross in the strategy library →
Strategies like this can run on your Blave Agent.
Browse ready-made strategies with real backtests →