The golden cross has an 86% win rate. It still loses to buy and hold.
Headline stats are not a backtest. Here's what the actual equity curve shows across 25 years.
The golden cross is the most popular systematic strategy on the internet. 50-day moving average crosses up over the 200-day, you buy. Crosses down, you sell. 79% win rate over 66 years. I ran the numbers from 2000 to today and got 86%. Profit factor of 20.
Looks incredible. Now look at the equity curve.
Buy and hold beats it by 50%. Same period. Same index. The strategy avoided the 2000 and 2008 crashes, which is the only reason anyone talks about it. Take those two events out and the drawdown matches the S&P, but the recovery is worse. Because in 2020 the death cross fired at the lows, and the golden cross fired right at the highs. Sold the bottom, bought the top. Same thing happened in the 2025 tariff tantrum.
Bear markets are faster now. The 50-200 cross is too slow to react. It made sense in 1960. In a market that draws down 30% in six weeks and recovers in three months, the slow averages are now the bug.
This is what a real backtest looks like. Not the headline win rate. The equity curve, the bear markets, the recovery periods, the chop zones. A 79% win rate on a strategy that loses to buy and hold is a stat designed to sell you something.
The premise isn’t broken. The parameters are. A 10-day or 5-day crossing the 200 might still work. Same logic, faster reaction, different market.
17.30% max drawdown over 25 years while the S&P drew down 57%. That came from systems that account for how fast bear markets are now. When those systems go into drawdown, members hear about it directly in The Drawdown Memo.
Free 25-Year Backtest PDF: https://www.statsedgetrading.com/the-25-year-backtest
Stats Edge Pro at $149.99/month with a 30-day money-back guarantee, or $1,499.99/year. Founding tier at $3,000/year for traders who want a 1:1 onboarding call.
— Michael Nauss, CMT, CAIA, CDMS