A golden cross is when the 50-day simple moving average climbs above the 200-day; a death cross is the reverse, and on the S&P 500 since 1950 neither has been a reliable timing signal on its own. Across 38 golden crosses and 38 death crosses, the golden cross beat an ordinary day's forward returns by a modest margin. The death cross mostly meant slower gains, not a crash, and it reversed within about three months 13 times. Here is the data, the arithmetic behind the lag, and where the index stood at the October 16, 2026 close.
What the golden cross and death cross actually are
Both signals come from two simple moving averages (SMAs) of closing prices. The 50-day SMA is the average of the last 50 closes. The 200-day SMA is the average of the last 200. StockCharts' ChartSchool notes that many chartists use the 50-day and 200-day together, and calls the 200-day the most popular long-term moving average.
- Golden cross: the 50-day SMA closes above the 200-day SMA after being below it.
- Death cross: the 50-day SMA closes below the 200-day SMA after being above it.
That is the whole definition. There is no volume filter, no confirmation rule and no magic. Everything below uses those exact rules on S&P 500 index closes from Financial Modeling Prep, January 3, 1950 through October 16, 2026: 19,223 sessions.
Why the cross always shows up late
ChartSchool puts it bluntly: moving averages are lagging indicators that will always be a step behind. The cross is slower still, because it waits for one lagging line to catch another.
The arithmetic is simple. The average close inside a 50-day SMA is about 24.5 sessions old. Inside a 200-day SMA it is about 99.5 sessions old, roughly five months. A single bad day barely moves either line. At the October 16, 2026 close of 7,659.44, a close 1% lower (about 77 points) shifts the 50-day SMA by about 1.5 points and the 200-day by about 0.4 points.
Run the numbers forward from that date. If the index fell a steady 0.5% every session, the death cross would arrive after 33 sessions, with the index already about 15% below the October 16 close. At a steady 1% a day it takes 25 sessions and a drop of about 22%. If the index simply went flat at 7,659.44, the death cross would not print for 135 sessions, purely because older, lower closes would roll out of the 200-day window. By the time the lines touch, a large part of the move has already happened.
Do they work? 76 years of crosses against the baseline
The common mistake is comparing returns after a cross to zero. Stocks rose most of the time over this period, so almost any signal looks good against zero. The honest test is against the baseline: forward returns from any day since October 18, 1950.
| Forward window | After golden (median) | After death (median) | Any day (median) | % positive: golden / death / any day |
|---|---|---|---|---|
| ~1 month (21 sessions) | +1.28% | -0.37% | +1.10% | 63% / 47% / 62% |
| ~3 months (63) | +4.52% | +0.20% | +2.73% | 76% / 53% / 67% |
| ~6 months (126) | +6.49% | +1.89% | +5.13% | 82% / 55% / 71% |
| ~12 months (252) | +13.10% | +10.54% | +10.47% | 82% / 71% / 74% |
Three things stand out. Golden crosses were followed by better-than-baseline medians at every horizon, but the edge at one month was under 0.2 percentage points. Death crosses were followed by weaker-than-baseline returns for about six months. At twelve months, the median after a death cross (+10.54%) was essentially the same as any random day (+10.47%).
So the death cross, on this sample, flagged a softer stretch, not a reliable crash warning. The sample is small (38 of each), the forward windows overlap, and the series is price only, with no dividends.
Whipsaws: the death cross cries wolf more often
A whipsaw here means the opposite cross arrived within 63 sessions, about three months. Death crosses whipsawed 13 times out of 38, or 34%. Golden crosses whipsawed 4 times out of 38, or 11%.
That asymmetry fits the timing. Death crosses print after prices have already fallen, and in 2020 and 2025 that was close to where the index turned back up. Anyone who sold on the signal and waited for a golden cross to buy back paid for the round trip.
The crosses since 2000 that teach the most
Twenty-four crosses printed from 2000 through October 16, 2026. These four cover the range:
| Date | Cross | S&P 500 close | Next ~3 months | Next ~12 months | What it shows |
|---|---|---|---|---|---|
| Dec 21, 2007 | Death | 1,484.46 | -9.7% | -41.3% | The signal that worked: ahead of the 2008 collapse |
| Mar 30, 2020 | Death | 2,626.65 | +16.2% | +50.7% | Printed one week after the March 23 low |
| Apr 14, 2025 | Death | 5,405.96 | +15.9% | +30.3% | Whipsaw: reversed 53 sessions later |
| Jul 1, 2025 | Golden | 6,198.02 | +7.9% | +20.7% | Still in force at the Oct 16, 2026 close |
The 2007 death cross came about 5% below the October 9, 2007 peak of 1,565.15, and the index still had about 54% further to fall to its March 9, 2009 close of 676.53. That is the case cross fans remember.
March 2020 is the mirror image. The index had already fallen about 22% from its February 19 peak and was about 17% above its March 23 low of 2,237.40 when the death cross printed. The April 2025 cross was similar: about 8% above the April 8, 2025 low, then reversed into the July 1 golden cross, which itself printed about 24% above that low.
Worked example: where the S&P 500 stands as of October 16, 2026
At the October 16, 2026 close:
- S&P 500 close: 7,659.44
- 50-day SMA: 7,706.42
- 200-day SMA: 7,271.44
- Gap between the averages: about 435 points, with the 50-day 5.98% above the 200-day
- Close versus 200-day: +5.34%. Close versus 50-day: about -0.6%
- Last cross: golden, July 1, 2025
The read: the long-term trend measure is still in golden-cross territory, while price has dipped just under its 50-day average. As the scenarios above show, closing a 435-point gap between the averages would take weeks of steady selling, so no death cross was close to printing on these numbers. For a broader read on how many individual stocks are above their own long-term average, see how many stocks are above their 200-day average.
The simpler 200-day rule, and what it costs
Many traders skip the cross and just compare price to the 200-day. We tested the plain version: hold the S&P 500 only when the prior close was above its 200-day SMA, from October 19, 1950 to October 16, 2026.
- Rule: 248.01x growth, invested 72% of sessions, 442 switches in or out.
- Buy and hold: 382.78x growth.
Annualized, that works out to roughly 7.5% a year for the rule against roughly 8.1% for buy and hold. The rule trailed, and the test leaves out things on both sides. It ignores trading costs and taxes on 442 switches, which would hurt the rule. It also ignores interest on cash while out of the market and dividends while in, so neither side gets its full real-world return. What the rule historically offered was time out of the market during long bear phases, not higher total growth.
Common mistakes with golden and death crosses
- Comparing to zero. Positive returns after a golden cross mean little when 74% of all 12-month stretches since 1950 were positive anyway.
- Treating the death cross as a crash signal. On this data its 12-month median matched an average day. 2007 was the exception, not the rule.
- Ignoring the lag. The averages hold roughly 24.5 and 99.5 sessions of old data. A cross confirms what already happened.
- Applying index results to single stocks. Everything here is the S&P 500 index. A cross on one stock is a different, untested signal.
- Forgetting costs. Whipsaws are round trips, and the backtests above charge nothing for them.
More technical breakdowns live in our Analysis section, and daily market coverage is on the Traders On Trend homepage.
Bull/Bear Verdict
Bull Case: Where it helps: the cross is a clean, rules-based trend filter. Golden crosses on the S&P 500 were followed by better-than-baseline median returns at every horizon since 1950, they rarely whipsawed (4 of 38), and the 2007 death cross flagged the 2008 bear market.
Bear Case: Where it misleads: the signal is built from months-old prices, so it lands late. Death crosses whipsawed 13 of 38 times, printed near the 2020 and 2025 lows, and their 12-month median matched any random day.
FAQ
Is a golden cross bullish?
Historically, mildly. On the S&P 500 since 1950, median returns after golden crosses beat the any-day baseline at 1, 3, 6 and 12 months, but the edge was small: +13.10% versus +10.47% at 12 months.
Does a death cross mean a crash is coming?
Not on this data. Returns after S&P 500 death crosses trailed the baseline for about six months, but the 12-month median was +10.54%, about the same as any day. The December 2007 cross before the 2008 drop is the notable exception.
Why use the 50-day and 200-day?
Convention. They are the pair many chartists watch, which is also why crosses make headlines. The same math works for any two lengths, and shorter averages react to new prices sooner because each close carries more weight.
When was the last S&P 500 golden cross?
July 1, 2025, at a close of 6,198.02. As of the October 16, 2026 close it was still in force, with the 50-day SMA at 7,706.42 above the 200-day at 7,271.44.