Moving average crossovers.

Put two moving averages on the same chart, one short and one long, and sooner or later they cross.

Traders read that moment as a change in direction: the short line moving above the long one is a buy signal, and moving below it is a sell signal.

The idea is simple to test and easy to automate. Its weakness is built in: both lines are made from past prices, so the cross always arrives after the turn it is meant to catch.

What is a moving average crossover?

A moving average crossover is the point where a faster moving average crosses a slower one. The fast line averages fewer periods, so it follows price closely. The slow line averages more periods, so it moves more calmly.

When the fast line crosses above the slow line, the latest prices are higher on average than the longer window. This is called a bullish crossover. When the fast line crosses below, it is a bearish crossover.

A simpler version uses one average and the price itself, with a close above the average as the signal. The two-line version reacts more slowly, and that trade-off runs through this whole lesson.

The previous lesson used moving averages to read the trend. A crossover turns that reading into a rule with a definite moment to enter.

Which averages do traders pair?

Any two periods can be paired, as long as one is shorter than the other. Some pairs come up often enough to have names.

You can build a crossover from a simple moving average or an exponential moving average. With the same periods, an EMA weights recent prices more, so EMA crossovers tend to come a little sooner and to flip more often.

The periods themselves carry no special power. 50 and 200 are round numbers that became a convention, and no pair of periods is known to be best for every market.

A crossover worked through, day by day

Here are 14 daily closes on EUR/USD. They are illustrative numbers chosen to keep the maths easy.

The fast line is a 3-day SMA and the slow line is a 6-day SMA. The mechanism is the same for 50 and 200.

Each average is the sum of the last N closes divided by N. On day 8, the 3-day SMA is (1.0950 + 1.0980 + 1.1010) ÷ 3 = 1.0980.

The 6-day SMA on day 8 adds the closes of days 3 to 8, which sum to 6.5850, and divides by 6 to get 1.0975.

Now apply a plain crossover rule to the table, using closing prices and ignoring the spread:

1. On day 7 the fast line, 1.09633, is still below the slow line, 1.09717. No signal.
2. On day 8 the fast line, 1.09800, is above the slow line, 1.09750. The lines have crossed, so the rule buys at the day 8 close of 1.1010.
3. On day 13 the fast line, 1.10200, drops below the slow line, 1.10283. The rule sells at the close of 1.0990.
4. The result is 1.0990 minus 1.1010, which is minus 0.0020, a loss of 20 pips.
5. On 0.10 lot of EUR/USD a pip is worth 10,000 × 0.0001 = $1, so the loss is $20 before the spread.

Why the crossover signal always lags

Look at where the trade sat inside the move. The lowest close was 1.0950 on day 6 and the highest was 1.1060 on day 10, a swing of 110 pips.

The buy came at 1.1010, 60 pips above the low. Price rose another 50 pips to 1.1060, then fell 70 pips to the exit at 1.0990. The rule caught none of the 110 pip swing and finished 20 pips down.

The cause is arithmetic: an N-period SMA averages the latest close and the N - 1 closes before it, so its centre sits (N - 1) ÷ 2 periods in the past.

The 3-day line therefore lags by about 1 day and the 6-day line by about 2.5 days.

For a cross to happen, price has to move far enough to drag the fast average past the slow one, and by then part of the move is already spent.

Longer pairs lag more. On a daily chart the 200-day SMA is centred about 100 trading days back, so a golden cross can confirm a rise that began months earlier.

Shorter periods cut the lag but cross more often on random noise. Longer periods filter the noise but enter and exit later. Changing the settings moves you along that trade-off and never removes it.

Where crossovers fail: sideways markets

The worked example was a short trend that reversed. In a market moving sideways, the two averages twist around each other and produce cross after cross, and each one can be a small loss plus the spread.

These repeated false signals are called whipsaws. A crossover rule cannot know in advance whether the market will trend or range, so it takes every signal and relies on an occasional long trend to pay for the small losses.

By construction, then, crossovers tend to look good on charts with long, smooth trends and poor on choppy ones. Results also shift with the pair, the periods, the timeframe and your trading costs.

Treat any claim that a particular pair of periods works as unproven until you have checked it yourself. The lesson on how to backtest a strategy shows how to run that test honestly, with costs included.

Ways traders try to filter crossover signals

Filters aim to skip some of the whipsaws. Each one also delays or drops some good signals, so every filter has a cost.

• Wait for the candle to close with the lines crossed, so a cross that reverses inside the bar does not count.
• Trade only in the direction of a third, slower average, for example taking a 20 and 50 buy signal only when price is above the 200.
• Require the slow average to slope in the direction of the signal.
• Check a trend-strength reading such as the ADX and ignore crosses when it is low.
• Require a minimum gap between the two lines, in pips, before treating the cross as real.

Exits are a separate decision. The worked example used the opposite cross to exit, which gave back 70 pips from the high. A rule can instead set a stop loss at entry and use the cross only to get in.

The MACD line is the gap between a 12 and a 26 period EMA, so its crossing zero is the same event as that EMA crossover.

Averages also act on price as levels, which is the subject of moving averages as dynamic support and resistance.