Golden cross trading: using the 50/200 cross as a trend filter

What the golden cross is

A moving average is the average closing price over a set number of bars. The 50-day simple moving average (SMA) is the average close of the last fifty trading days, roughly two and a half months. The 200-day SMA covers about ten months. Each day the oldest close drops out and the newest one comes in, so the line moves slowly with the market.

Because the 200-day line averages four times as many days, it moves far more slowly than the 50-day. In a long decline, both lines fall and the 50 sits below the 200, because recent prices are lower than older ones. When the market turns and rises for long enough, recent prices pull the 50-day line up faster than the 200-day line, and eventually the 50 crosses above. That is the golden cross. The death cross is the same event in reverse at the end of an advance.

The cross itself carries no new information. It is arithmetic: the average of the last fifty closes has risen above the average of the last two hundred. What it does is summarise, in one unambiguous event, that the medium-term trend has turned up relative to the long-term one. Many traders, funds and commentators watch it for that reason, which is why financial news reports it on indices, gold and the major pairs.

SMA or EMA, 50/200 or something faster

The classic definition uses simple averages on the daily chart. Some traders use exponential averages, which weight recent prices more and so cross a few days earlier, at the price of more false crosses. On lower timeframes the same idea appears as a 20/50 or 9/21 crossover; those are covered in the academy lesson on moving average crossover strategies. This page sticks to the 50 and 200-day SMAs, the version most people mean by the name.

Why the golden cross is always late

The honest thing to say about the golden cross is that it arrives late. For the 50-day average to rise above the 200-day, price has to have been climbing for weeks, often months. By the day of the cross the market is usually well off its low, and sometimes the first leg of the new trend is already over.

This is not a flaw that a better setting fixes. Lag is the price of smoothing: an average slow enough to ignore a week of noise is, by the same token, slow to react to a real turn. A faster pair of averages crosses sooner and gives more false signals; a slower pair gives fewer false signals and crosses later. There is no setting that is both early and reliable, which is the general point of the academy lesson on indicator lag and whipsaws.

So buying the day of the cross means buying after a long rise, often just before the pullback that rises usually take. A better use is to treat the cross as a change of regime. Before it, the long-term picture says down, and longs are against the tide. After it, the picture says up, and the job is to find a good place to join.

The strategy: cross, pullback, trigger

This version uses the golden cross as a filter and takes entries on pullbacks, which puts the stop closer and the entry price lower than buying the cross itself.

1. The filter

The 50-day SMA has crossed above the 200-day SMA, and both are now rising or the 200 has at least stopped falling. Price is above both lines. If the 200-day is still falling steeply, the cross is more likely to be a bounce inside a downtrend than a new trend, and it is worth waiting.

2. The pullback

Wait for price to fall back towards the 50-day average. In a healthy new uptrend, the first pullback after a golden cross often comes back to the 50, sometimes to the zone between the 50 and the 200. A pullback that also lands on an old resistance level that has turned into support is the strongest version.

3. The trigger

Enter when a daily candle closes back up after touching or nearing the 50-day line, for example a bullish engulfing candle or a candle with a long lower wick. Traders who want more detail drop to the 4-hour chart and wait for a break of the pullback's last lower high.

In the chart above, the market bottoms, climbs for weeks, and only then does the 50-day average cross the 200-day. Buying at the cross would have meant buying near a short-term high. The entry instead comes on the pullback to the 50-day average, with the stop under the pullback's low.

Stops, exits and position size

The stop goes below the low of the pullback that produced the entry, plus a buffer for the spread and for wicks. On a daily EUR/USD chart that might be 60 to 120 pips; on gold, $30 to $60. Putting the stop below the 200-day average instead sounds logical but is often far too wide to size sensibly.

A daily stop is wide, so the position has to be small. Fix the risk per trade first, for example 1% of the account, and let the position size calculator work out the lot size from the stop distance. A 100-pip stop with $100 at risk is a tenth of the size of a 10-pip stop with the same $100.

Exits for a long-term trend strategy need room. Three common choices:

• A close below the 200-day average. Lets the trend run for months and accepts a large giveback at the end.
• The next death cross. Slower still, and it can return much of the gain, but it is the purest version of the strategy and the easiest to test.
• A trailing stop below swing lows. Move the stop up under each new higher low on the daily chart. A middle path between the two above.

A position held for weeks pays or earns swap every night. On some pairs that cost adds up, so include it in your testing.

When the golden cross fails: whipsaws

The golden cross works when a market moves from one long trend into another. It fails when the market goes sideways for months. In a range, both averages flatten and wind around each other, and every swing pushes the faster line across the slower one. A golden cross is followed by a death cross a few weeks later, then another golden cross, and a trader who acts on each one pays the spread and a stop every time.

There are a few filters that help, though none removes the problem:

• Slope. Ignore a cross when the 200-day average is flat. A real trend change bends the slow line too.
• Distance. When the two averages have been within a small distance of each other for weeks, the market is ranging, and the next cross means little.
• Structure. Check whether price is making higher highs and higher lows, as in the swing trading strategy. A cross without a trend structure is a coincidence of averages.

Accept that some whipsaws will get through any filter. The strategy depends on the occasional long trend paying for the false starts, so the losses on whipsaws have to stay small.

How to test a golden cross strategy

Golden crosses are rare, a few per decade per market on the daily chart, so testing one market is not enough. Use several: EUR/USD, GBP/USD, USD/JPY, gold and a stock index. Go back as far as your platform allows.

1. Mark every golden cross and death cross on each chart.
2. For each golden cross, apply the pullback rules and record the entry, stop and exit the rules produce. Record the crosses that never gave a qualifying pullback too.
3. Compare with the simplest alternative, buying on the cross and selling on the death cross, so you know whether the pullback rule adds anything.
4. Include the swap for every night held and the spread on entry and exit.

Expect long gaps between trades and long stretches of losing whipsaws. The numbers to study are the average win against the average loss and the longest run of losers, then put them through the risk of ruin calculator. A strategy that trades a few times a year needs a risk per trade you can live with through several years of results.