The golden cross strategy is the most famous signal in technical analysis: buy when the 50-day moving average crosses above the 200-day, sell on the "death cross" when it falls back below. It is also one of the laziest to backtest badly. We ran the 50/200 moving average crossover and the faster 20/50 crossover on daily Nasdaq-100 futures (NQ) from June 7, 2010 to April 16, 2026, and compared both with simply holding.
Where the golden cross comes from
Nobody owns the golden cross. Moving average crossovers have been part of market lore for most of the last century, and the 50-day and 200-day pairing became the version quoted on financial television. We found no reliable single inventor, so we won't name one. The 20/50 crossover is a common faster variant used by swing traders who find the 50/200 too slow.
Exact rules of the golden cross strategy we tested
- Averages: simple moving averages (SMA) of the daily close.
- Entry (golden cross): the fast SMA (50 or 20) closes above the slow SMA (200 or 50). Buy at that close.
- Exit (death cross): the fast SMA closes back below the slow SMA. Sell at that close.
- Long only. No stop-loss, no target, no filters.
- Costs: 1.0 NQ point per round trip.
All results are in NQ index points per 1 contract. A Micro E-mini Nasdaq-100 (MNQ) contract is worth $2 per point and a full NQ contract $20 per point, so MNQ dollars below are simply points × 2. Every trade has 1.0 point deducted per round trip to cover slippage and commission.
50/200 moving average crossover vs 20/50 vs buy and hold
| Metric | 50/200 golden cross | 20/50 cross | Buy & hold |
|---|---|---|---|
| Trades | 13 | 45 | — |
| Win rate | 61.5% | 60.0% | — |
| Avg trade (pts) | +1,346.65 | +296.68 | — |
| Total (pts) | +17,506.5 | +13,350.5 | +24,677.2 |
| Profit factor | 26.85 | 2.9 | — |
| Closed-trade drawdown (pts) | 384.5 | 2,791.5 | 5,878.2 (mark-to-market) |
| Avg win / avg loss (pts) | 2,272.97 / -135.45 | 754.97 / -390.76 | — |
| MNQ total ($, 1 contract) | $35,013 | $26,701 | $49,354 |
| Points 2010–2019 | +3,991.8 | +4,130.2 | — |
| Points 2020–2026 | +13,514.8 | +9,220.2 | — |
| Profit factor 2020–2026 | n/a (no losing trades) | 2.7 | — |
The 50/200 golden cross strategy made 17,506.5 points on 13 trades, about 70.9% of buy and hold's 24,677.2 points while sitting in cash through each death cross. The 20/50 crossover made 13,350.5 points on 45 trades, about 54.1% of buy and hold.
Why 13 trades is not a statistic
The 50/200 profit factor of 26.85 looks spectacular. Ignore it. Thirteen trades over almost 16 years is not a statistically meaningful sample. Eight winners and five small losers (the average loss was -135.45 points) produced that number, and most of the profit came from a handful of long rides: 4,760.5 points from the trade entered in 2020 and 6,559.5 from the one entered in 2023. Between 2020 and 2026 there was no losing closed trade at all, which is why the recent-era profit factor can't even be computed. Change the start date by a few months, or lose one of those big trends, and the headline figures move a lot.
The 20/50 crossover's 45 trades are a bigger sample, though still under what we'd like. Its profit factor of 2.9 held up in the second half (2.7 in 2020–2026), and it had a real bad year: -2,568.2 points in 2022.
The drawdown trap: don't compare these two numbers
The 50/200 closed-trade drawdown of 384.5 points sits next to buy and hold's 5,878.2 points in the table. It is tempting to conclude the golden cross cut drawdown by more than 90%. It did not, or at least this data can't show that. The two figures measure different things:
- Buy and hold drawdown is the worst peak-to-trough fall of the daily closing price, marked to market every day.
- Closed-trade drawdown only looks at the equity curve after each trade closes. While a golden-cross position is open, it can give back thousands of points from its peak before the death cross finally fires, and none of that appears in the 384.5-point figure.
Because the 50/200 holds for months at a time, its open-trade giveback before an exit can be large. A fair comparison would need a daily mark-to-market equity curve for the crossover, which this test did not produce. What we can say: the golden cross did sidestep the long stretches below the 200-day average, and it never closed a large losing trade.
Robustness: 2010–2019 vs 2020–2026
The 50/200 made 3,991.8 points in 2010–2019 and 13,514.8 in 2020–2026. The 20/50 made 4,130.2 and 9,220.2. Both were profitable in both halves, and point totals naturally grew as NQ's price rose from roughly 1,800 toward 26,000.
Our take: On a market that went up this much, a golden cross strategy is best thought of as buy and hold with an emergency exit. It gave up roughly 29% of the points in exchange for being flat through the worst trends down, and the 20/50 gave up more because it got whipsawed more. If your goal is maximum points, holding won. If your goal is avoiding a sustained bear market with a simple rule, the 50/200 did its job here, but with just 13 trades we can't call it proven. Don't trade it for its profit factor.
For a faster trend-following approach on the same data, see our 9/20 EMA pullback test and the Weinstein Stage 2 weekly backtest.
What this backtest does not model
- Roll gaps. The test runs on a continuous front-month NQ series, so quarterly contract rolls can create small artificial price jumps.
- Perfect fills. Signals are computed on the daily close and filled at that same close. In real life you need to act in the last minutes of the session or accept the next open.
- Overnight and weekend risk is in the price series, but position sizing, margin calls and psychological stamina through deep open drawdowns are not.
- No taxes or financing costs are modeled beyond the 1-point round-trip deduction.
This is a backtest on one instrument. Past results do not guarantee future results, and nothing here is financial advice.
Key takeaways
- 50/200 golden cross: 17,506.5 points on 13 trades, about 71% of buy and hold (24,677.2).
- 20/50 crossover: 13,350.5 points on 45 trades, PF 2.9.
- Thirteen trades is too small a sample to trust the 50/200's 26.85 profit factor.
- Closed-trade drawdown (384.5) is not comparable to buy-and-hold drawdown (5,878.2).
- Both crossovers made money in both halves of the test.
Put the research to work on your Desk
Explore the tools behind the research, compare the available access, or open your Desk if you already have an account.
FAQ
Does the golden cross strategy work on Nasdaq futures?
The 50/200 golden cross made 17,506.5 NQ points from 2010 to 2026 on only 13 trades, versus 24,677.2 for buy and hold. It was profitable, but 13 trades is not a statistically meaningful sample.
Is the 20/50 crossover better than 50/200?
Not on NQ. The 20/50 made 13,350.5 points on 45 trades with a 2.9 profit factor, less than the 50/200. It trades more and gets whipsawed more.
Did the golden cross reduce drawdown versus buy and hold?
We can't say from this test. The golden cross figure is closed-trade drawdown, which ignores losses while a position is open; buy and hold's figure is marked to market daily. The two aren't comparable.
What is a death cross?
The death cross is the opposite of the golden cross: the 50-day moving average closing below the 200-day. In our test it was the exit signal.
Loomi.Ai provides educational market research, not financial advice. Trading futures and options involves substantial risk of loss, and backtested results are historical and do not guarantee future returns.
Sources
- Loomi backtest: blog_strategy_tests.py + results JSON, Loomi research