Death cross vs golden cross: what they actually signal (and what they don't)

Both "death cross" and "golden cross" get more airtime in financial media than they deserve. CNBC mentions them, retail Twitter quotes them, charting sites pop alerts on them. But almost nobody actually defines them carefully or asks whether the historical track record justifies the attention.

The short version: they're lagging regime labels, not entry/exit signals, and the SPY data over four decades shows the actual cross events are much rarer and much noisier than the coverage suggests. Here's what they are, what they aren't, and why the SMA200 framework focuses on something different.

Definitions, exactly

The two events get confused with a related-but-different signal that happens far more often. Worth being precise:

Golden cross: the 50-day simple moving average of a security crosses ABOVE the 200-day simple moving average.

Death cross: the 50-day SMA crosses BELOW the 200-day SMA.

Neither one is about price. Both are about the relationship between two smoothed averages of price. The 50-day SMA is itself a lagging indicator, smoothed over ~10 weeks; the 200-day SMA is smoothed over ~10 months. A cross between them is the slow part of price action catching up to the slower part.

What gets confused with these:

Price crosses its own 200-day SMA: the daily close of a security moves from below to above (or above to below) the 200-day SMA itself. This is what most people mean when they casually say "death cross" or "the SMA200 was broken." It is NOT a death cross technically. It's the underlying event the SMA200 framework actually tracks.

The distinction matters because these two events happen at very different frequencies.

Frequency on SPY since 1985

Approximate cross counts over the ~40-year SPY (and its index predecessors) record:

• Golden crosses: roughly 6 to 8 times
• Death crosses: roughly 6 to 8 times
• Price-crosses-its-own-SMA200 events: roughly 50 to 70 times

So the 50/200 cross events happen about once every 5 to 6 years on average. Price-vs-SMA200 events happen about once every 7 to 8 months on average. Order-of-magnitude different in frequency.

That gap is exactly what makes 50/200 crosses get media coverage: they're rare enough to feel meaningful when they happen. It's also what makes them mostly useless as trading signals: by the time the 50-day catches up to the 200-day, the move has already played out for months.

The famous false signals

Three crosses everyone in this space points at, with rough dates:

The 2015 to 2016 SPY whipsaw. A death cross fired in August 2015 after a brief mid-summer correction. By April 2016 SPY had recovered, the 50-day caught back up, and a golden cross fired. Anyone who acted on the death cross sold the bottom and bought back higher. Six months round trip, multi-percent drawdown on the round trip itself.

The 2022 SPY death cross. Fired in February 2022, after the January selloff. SPY was ~$435. SPY then continued lower to ~$360 in October before reversing. The cross "worked" in that it correctly labeled the bear regime, but it also fired late (after most of the January damage was done) and gave back ground in the chop. Not the clean signal it gets quoted as.

The 1990 SPY golden cross. Fired in late 1991 after the recession. By the time the cross confirmed, SPY had already rallied substantially off the lows. Anyone waiting for the cross to enter missed roughly the first 20% of the bull market.

The pattern these share: 50/200 crosses are lagging signals derived from already-lagging signals. They confirm regimes that have already started. They never call tops or bottoms; by construction they can't.

What they signal in practice

The 50/200 cross is best understood as a regime confirmation label, not a trade trigger:

• A golden cross says "the long-term uptrend has confirmed itself by averaging out across the last ~10 weeks AND the last ~10 months."
• A death cross says the same in reverse.

It's a high-confidence-low-timeliness reading. Useful for narrative ("we're in a confirmed bear regime") but rarely useful for action (the action window already passed).

The SMA200 framework treats the more frequent event (price crosses its own SMA200) as the actual signal worth watching, because:

1. It's an order of magnitude more frequent, so it actually generates trade-relevant events
2. It catches regime changes within weeks rather than within months
3. It can be alerted-on cleanly without conflating two different smoothing horizons

What they don't signal

A clean list of what 50/200 crosses are NOT:

• Not predictions of future returns. They label a regime that has already started.
• Not high-confidence at all when they fire. The 2015 and 2022 examples above are recent cases where the cross fired and the subsequent move went the "wrong" direction for the cross direction.
• Not better than price-vs-SMA200. The simpler signal (just the underlying's daily close compared to its own 200-day SMA) catches the same regime changes earlier and with comparable noise.
• Not exclusive to equities. The same math applies to any time series with enough history; the equity-specific framing in financial media is convention, not necessity.

What to actually watch

The framework on sma200.trade tracks the simpler event: the daily close of a security vs its own 200-day SMA. That signal:

• Fires more often than the 50/200 cross (useful for trade decisions on a reasonable timeframe)
• Catches regime changes earlier (~3 months earlier on average than the corresponding 50/200 cross)
• Has the same long-run "trend confirmation" property without the second moving-average lag stacked on top

The framework treats "stock is above its 200-day SMA = long-term uptrend" and "below = downtrend" as the binary read. That's the signal alerts and the daily refresh are built around. Not the 50/200 cross, which is a downstream artifact of the same underlying trend.

Practical takeaway

When financial media coverage spikes around a death cross or golden cross firing, the right interpretation is: the regime has been changing for weeks or months already. The cross is the late-stage label, not the inflection.

The underlying signal (price vs SMA200) has already been alerting on this regime change for some time. Anyone tracking that signal directly has been ahead of the news cycle by months.

If you want one email when YOUR ticker crosses its own 200-day SMA (the simpler, earlier signal), the per-ticker pages at sma200.trade take 30 seconds to set up. No 50/200 cross calculations, no media-narrative lag. Just the underlying event.