On Natural Gas, the Golden Cross Is a Sell Signal

The death cross and the golden cross, a market's 50-day average crossing below or above its 200-day, are the most watched trend signals on any chart. They are built to do one job: confirm that a trend has taken hold. Natural gas is close to the worst market on earth to ask that of. It is the most mean-reverting of the major commodities, nicknamed the widow maker for how violently it snaps back on anyone who chases it. We took every 50/200-day cross on Henry Hub natural gas futures since 2000, 21 golden and 21 death crosses, and tracked the price before and after each one. The signals fire late, and more often than not they point the wrong way.

The golden cross buys the top

A golden cross is supposed to be the all-clear, the moment a downtrend gives way to an uptrend. On natural gas it has been close to the opposite. In the month after the 21 golden crosses since 2000, gas fell a median 5.4%, and finished higher just 29% of the time. A coin flip from any random day is 50%. The slow cross only certifies an uptrend after gas has already run, and by then the run is usually stretched. Mean reversion takes over and drags the price down for roughly two months before it steadies, and only by the six-month mark does it drift back into the green.

Forward return on Henry Hub natural gas futures after each signal, versus the return from any random day (the baseline every signal has to beat). 2000 to 2026.

Horizon	After a golden cross	After a death cross	Any day (baseline)
1 month	-5.4% (29% up)	-0.5% (48% up)	+0.0% (50% up)
2 months	-5.8% (43% up)	+0.3% (52% up)	-0.1% (50% up)
3 months	-0.6% (48% up)	-0.1% (50% up)	+0.4% (51% up)
6 months	+5.2% (52% up)	-2.5% (50% up)	+1.5% (52% up)

And this is not just the calendar at work. When we strip out seasonality, comparing each cross to the typical move for that same month, the near-term drop barely budges (a 5.1% shortfall at one month). The golden cross is buying the top. December 2019: a golden cross at $2.29, then down 24% in two months. October 2023: a golden cross at $3.34, then down 47% in six. November 2025: a golden cross at $4.37, then down 31% in two months.

The death cross often marks the bottom

If the golden cross buys the top, the death cross has a habit of selling the bottom. After the 21 death crosses, gas was essentially flat over the next month (a median of -0.5%, near a coin flip), and several of the most violent rallies in the data began at one. Gas printed a death cross at $1.96 in August 2024 and then climbed 105% over the next six months. It death-crossed in February 2022, a few weeks before war doubled the price, up 99% in sixty days. The famous sell signal kept firing into lows.

A spread of individual crosses: the price at the signal, and what natural gas did next. Death crosses repeatedly landed near lows; golden crosses near highs.

Signal	When	Gas price	Next 60 days	Next 6 months
Death cross	Aug 2024	$1.96	+53%	+105%
Death cross	Feb 2022	$4.23	+99%	+96%
Golden cross	Dec 2019	$2.29	-24%	-21%
Golden cross	Oct 2023	$3.34	-23%	-47%
Golden cross	Nov 2025	$4.37	-31%	-34%
Death cross	Nov 2022	$6.24	-62%	-66%

That does not make the death cross a buy signal either. Its median six-month drift is mildly negative, and one death cross, November 2022 at $6.24, correctly caught the crash off the war spike, down 66% in six months. The honest read is narrower: the death cross is simply not the warning of lower prices that its name implies. The average forward return after one is actually positive, because those rebounds are so large.

Even the trend does not trend

It is not just the cross as an event. We split every single trading day by whether gas was in an uptrend (the 50-day above the 200-day) or a downtrend, and measured what came next. The uptrend never beat the downtrend. At six months the downtrend actually did better, a median +4.6% against -0.7%.

Forward return on every day, split by whether natural gas was above or below its own trend. Being in an uptrend did not help.

Horizon	In an uptrend (50 above 200)	In a downtrend (50 below 200)
1 month	+0.5%	-0.3%
3 months	-0.2%	+0.6%
6 months	-0.7%	+4.6%

On natural gas, sitting above the 200-day average tells you nothing good about the next six months. The market reverts faster than a 200-day average can ever confirm a trend.

What the crosses really track

So if not direction, what do the crosses mark? The seasons. They are not scattered randomly through the year. Golden crosses bunch into October, November, and December, as gas climbs into winter heating demand. Death crosses bunch into July and August, the shoulder-season lows. The 50/200 cross on natural gas is, more than anything, a slow seasonal clock, and that seasonal pattern is something you can see directly in our BOIL winter study. The cross adds no directional edge on top of the calendar.

On the futures price the crosses are a wash. On a fund you can actually buy, the deck is tilted against you before you start. UNG, the main natural gas ETF, fell a median 14% over any six-month window in our data, because holding and rolling futures in a market that is usually in contango is a slow, constant leak. A golden cross occasionally caught one of gas's rare sustained up-phases and so beat that grim baseline, but with only ten events and a six-month edge that faded to under 2%, it is not a timing tool you can lean on. The trend filter did not plug the leak either.

How we tested it

• Data. Henry Hub front-month natural gas futures, daily from 2000 to today, about 6,500 trading days.
• The signals. A golden cross is the day the 50-day simple moving average closes above the 200-day; a death cross is the day it closes below. Crosses within 20 trading days of each other are counted once. That gave 21 of each.
• Forward returns. The price change 20, 40, 60, and 120 trading days (about one to six months) after each cross, measured against an all-days baseline so every signal has a fair bar to clear.
• Seasonality control. Each cross compared with the median forward move for that same calendar month, to separate the signal from natural gas's strong seasonal pattern.
• Trend state. The same forward returns computed for every day, split by whether the 50-day was above or below the 200-day.
• Robustness. The whole test repeated on UNG, the investable natural gas ETF, from 2007. Front-month futures carry roll gaps, while UNG reflects the roll-inclusive return an investor actually earns.