Does the Golden Cross Work? A 25-Year Backtest on US, Canadian and UK Stocks
By MarketInOut · October 26, 2026
The golden cross is one of the best-known signals in technical analysis: when a stock's 50-day moving average crosses above its 200-day average, the trend has turned up. The death cross, the 50-day falling back below the 200-day, says it has turned down. Financial news covers both every time they appear on a major index.
But does trading it make money? We backtested the golden cross on the members of three large-cap indexes, the S&P 500, the S&P/TSX Composite and the FTSE 100, and compared it with two things: a local index fund bought and held, and a simpler trend rule, buying when the price crosses above its 200-day average. The answer was different in each country.
The short version
In Canada, the golden cross on TSX Composite members made 10.5% a year from 2001 to October 2026, against 8.4% for an S&P/TSX 60 index fund (XIU). It stayed ahead without the strong 2025–26 run (9.3% vs 7.2%) and since 2010 (11.2% vs 9.8%).
In the US, it trailed the market: 7.1% a year on S&P 500 members vs 9.0% for SPY. Since 2010 the gap was wider: 7.9% vs 13.9%.
In the UK, it made 6.1% a year vs 4.8% for a FTSE 100 fund since 2007, but the edge came from 2008. Since 2010 it trailed: 5.3% vs 6.0%.
It beat the plain 200-day rule in all three markets over the full period, with a fifth to a sixth of the trades: 673 trades vs 3,715 on S&P 500 members.
In every calendar year the index fund fell, the golden cross portfolio fell less or gained, in all three markets. It lost 33% in 2008 on US stocks, against 36% for SPY.
Annual return by market
Average annual return of each strategy over the full test period:
Full results
Market
Since
Golden cross
Above 200-day
Index fund
S&P 500 members
2001
7.1% / 44%
5.6% / 59%
SPDR S&P 500 (SPY): 9.0% / 54%
S&P/TSX Composite members
2001
10.5% / 39%
5.5% / 44%
iShares S&P/TSX 60 (XIU): 8.4% / 47%
FTSE 100 members
2007
6.1% / 32%
4.5% / 24%
iShares Core FTSE 100 (ISF): 4.8% / 44%
Annual return / maximum drawdown. Each test ends on October 23, 2026. The FTSE 100 test starts in 2007, the first year with index membership history for it. Index funds are bought and held, with dividends reinvested. Results are in each market's own currency.
How the trades looked:
Market
Rule
Trades
Winners
Average hold
S&P 500 members
Golden cross
673
43%
186 days
S&P 500 members
Above 200-day
3,715
17%
34 days
TSX Composite members
Golden cross
618
44%
196 days
TSX Composite members
Above 200-day
4,080
15%
31 days
FTSE 100 members
Golden cross
443
40%
176 days
FTSE 100 members
Above 200-day
2,756
15%
30 days
Is one period driving it?
A long backtest can hide one great stretch. So we ran the same strategies again from 2010, and also took the last two years out of the full test:
Market
Golden cross since 2010
Above 200-day since 2010
Index fund since 2010
Golden cross, to end of 2024
Index fund, to end of 2024
S&P 500 members
7.9%
10.4%
13.9%
7.2%
8.4%
TSX Composite members
11.2%
5.4%
9.8%
9.3%
7.2%
FTSE 100 members
5.3%
5.3%
6.0%
5.5%
3.4%
Canada held up in every version. The UK did not: most of its edge came from 2008, when the strategy lost 9% and the FTSE 100 fund lost 28%. From 2010 on, the fund was ahead. In the US, the golden cross never caught up with SPY, and since 2010 even the simpler 200-day rule did better.
We also changed the portfolio from 20 positions at 5% each to 50 positions at 2% each. The US result improved to 8.5% a year, still below SPY. Canada stayed ahead at 9.5% a year. The UK fell to 3.7% a year, below the fund, so the UK result also depends on how the portfolio is built.
What about the golden cross on the index fund itself?
Many investors use the signal on the index, not on single stocks: hold the fund while its 50-day average is above its 200-day, and sell on the death cross. We tested that too. It made very few trades (13 to 18 in each market) and cut the worst drawdown everywhere, but it lowered the return everywhere as well:
Fund
Golden cross timing
Buy and hold
SPY, from 2001
8.3% / 33%
9.0% / 54%
XIU, from 2001
4.8% / 38%
8.4% / 47%
ISF, from 2007
2.2% / 37%
4.8% / 44%
Annual return / maximum drawdown. Cash earned no interest while out of the market, which makes timing look a little worse than it would be with cash in a savings account or T-bills.
In the US, timing cost 0.7 points a year and saved a lot of pain. In Canada and the UK it cost far more. In those markets the death cross often came late in a fall, and the next golden cross came only after much of the rebound.
What this means
The golden cross is a slow trend filter, not a timing tool. It held stocks for about six months on average and was right less than half the time. Its returns came from a few long trends, and its value showed most in bear markets, when it was out of many falling stocks.
Slow beat fast. Buying every stock that crossed above its 200-day average made five to seven times as many trades, and only 15–17% of them were winners. The 50-day/200-day pair filtered out many of these false starts.
The market matters. On index members, the same rules beat the index fund in Canada in every version we ran, and lost to it in the US in every version. We didn't test why. The largest US stocks have been hard to beat with any rule since 2010, because buy-and-hold already owned the winners.
Smaller losses don't mean a smaller risk of loss. The strategy still fell 33% in 2008 on US stocks and 27% on Canadian stocks. The signal reacts to price, so a fast crash hurts before the death cross arrives.
Limits of this test
We didn't rank the signals. When more stocks crossed on one day than there were free positions, the backtester's default order decided which ones were bought. A ranking rule, such as relative strength, would change the results.
Three markets and a few variations are a small sample. A strategy that beats its benchmark in one of three markets can do so by chance.
Taxes are not included. The golden cross trades far less than the 200-day rule, but far more than buy-and-hold.
The test uses closing prices for entries and exits, 0.1% commission per trade and a small spread. Real fills can differ.
How we tested
Golden cross: buy an index member on the day its 50-day simple moving average crosses above its 200-day average; sell when the 50-day crosses back below (the death cross). No stop-loss.
Above 200-day: buy when the price crosses above its 200-day average; sell when it closes below it.
Positions: up to 20 at a time, 5% of capital each, bought and sold at the closing price.
Costs: 0.1% commission per trade plus a small spread.
Data: prices adjusted for dividends and splits, delisted stocks included, and index membership as of each date, so the strategy only bought stocks that were in the index at the time.
Benchmarks: each index fund run as a one-ticker buy-and-hold backtest on the same dates.
Run it yourself
Opening formula, using S&P 500 members as the example:
index(sp500) and sma(50) ca sma(200)
Closing formula:
sma(50) cb sma(200)
For the 200-day rule, use index(sp500) and price ca sma(200) to open and price cb sma(200) to close. To time an index fund, open with ticker(spy) and sma(50) > sma(200) and use the same closing formula. Paste them into the Strategy Backtester and change the universe to index(tsx) or index(ftse100) for the other markets.
To see which stocks gave a signal today, open the Golden Cross screen for the US, Canada or the UK, and the Death Cross screen for the sell signals. The Above 200-Day MA screen lists the stocks in a long-term uptrend.
