MACD is a trend and momentum indicator built from moving averages. Its full name is Moving Average Convergence Divergence.
The default version compares a 12-period exponential moving average with a 26-period exponential moving average. A third average, called the signal line, tracks that difference. The histogram shows the distance between the two MACD lines.
That sounds more complicated than it is. The indicator answers three practical questions:
- Is the faster average above or below the slower one?
- Is the gap between them growing or shrinking?
- Has short-term momentum changed direction?
MACD does not tell you whether an asset is expensive or cheap. It also has no overbought or oversold boundary. I read it as a compact view of trend and momentum, not as a prediction of the next candle.
How to read MACD at a glance
| What you see | What it means |
|---|---|
| MACD line above zero | The 12-period EMA is above the 26-period EMA. |
| MACD line below zero | The 12-period EMA is below the 26-period EMA. |
| MACD above the signal line | The histogram is positive. Recent momentum has turned upward relative to its own average. |
| MACD below the signal line | The histogram is negative. Recent momentum has turned downward relative to its own average. |
| Histogram bars growing | The distance between MACD and its signal line is increasing. |
| Histogram bars shrinking | The distance is decreasing. A crossover may follow, but price has not necessarily reversed. |
The most common mistake is counting the line crossover and the histogram crossing zero as two confirmations. They are the same event shown in two ways. When MACD moves above its signal line, the histogram must become positive because the histogram equals MACD minus the signal line.
The three parts of MACD
MACD line
The MACD line is the 12-period EMA minus the 26-period EMA.
When the 12 EMA rises above the 26 EMA, MACD moves above zero. When it falls below the 26 EMA, MACD moves below zero. The farther the averages move apart, the farther MACD travels from the centerline.
If exponential moving averages are new to you, the moving average guide explains why an EMA reacts faster than a simple moving average.
Signal line
The signal line is usually a 9-period EMA of the MACD line. It smooths MACD itself, not price.
MACD reacts first. The signal line follows. Their crossover tells me that the recent direction of the MACD line has changed relative to its own nine-period average.
Histogram
The histogram is MACD minus the signal line.
Positive bars mean MACD is above the signal line. Negative bars mean it is below. Taller bars show a wider gap. Bars moving toward zero show that the gap is closing.
A shrinking positive histogram does not automatically mean bearish momentum. It means bullish momentum, as measured by this calculation, is increasing more slowly or starting to weaken. Price may pause, pull back or keep rising at a slower pace.
How MACD looked when gold moved into a trend
The chart below uses daily closes for continuous gold futures. It shows the 12 EMA, 26 EMA, MACD line, signal line and histogram calculated with the default 12, 26 and 9 settings.
MACD crossed above its signal line on February 23, 2024, while both lines were still below zero. The histogram turned positive on the same candle. Four days later, on February 27, MACD moved above zero as the 12 EMA crossed above the 26 EMA.
Those two events answer different questions:
- February 23: momentum improved relative to the signal line.
- February 27: the faster price average moved above the slower average.
The first event is faster but more vulnerable to a failed turn. The zero-line cross is slower because price has to move far enough to reverse the order of the two EMAs.
Gold closed at 2,038.60 on February 23 and at 2,382.30 on April 18, a rise of about 16.9%. I would not use that result to claim every bullish MACD crossover works. The useful part of the example is seeing what each line was measuring as the move developed.
Why MACD crossovers become noisy in a range
MACD is built from averages. When price keeps changing direction without going anywhere, the fast average repeatedly moves above and below the slower calculations.
From June 6 to August 15, 2025, continuous gold futures produced seven MACD signal-line crossovers. The closing price moved from 3,322.70 to 3,336.00, a net change of only 0.4%.
Price still moved inside the range. A trader with separate rules for entries, exits and costs might have found tradable swings. But blindly switching direction at every crossover would have meant frequent decisions without a sustained trend.
Before using a crossover, I check price first:
- Is price making higher highs and higher lows, lower highs and lower lows, or neither?
- Is MACD holding mostly on one side of zero?
- Did the crossover happen after a pullback in an existing trend or in the middle of a range?
- Is there a price level that invalidates the idea?
If price structure is mixed and MACD keeps crossing near zero, I treat the indicator as noise until the market chooses a direction.
Signal-line crossover versus zero-line crossover
These signals are often grouped together, but they have different speeds.
Signal-line crossover
A bullish crossover occurs when MACD moves above the signal line. A bearish crossover occurs when it moves below.
This is the faster signal. It can appear while MACD is still below zero in a downtrend or above zero in an uptrend. That makes it useful for spotting a change in momentum, but it can also fire during an ordinary pullback.
Zero-line crossover
MACD crosses above zero when the 12 EMA crosses above the 26 EMA. It crosses below zero when the order reverses.
This is slower. By the time it appears, price may already have moved a meaningful distance. I use it more as a trend filter than as a precise entry trigger.
A simple way to combine the two is to separate context from timing. The zero line describes the broader MACD state. The signal line shows the shorter shift inside that state. This does not remove false signals, but it prevents me from treating every crossover as identical.
What MACD divergence means
Bullish divergence appears when price makes a lower low while MACD makes a higher low. Bearish divergence appears when price makes a higher high while MACD makes a lower high.
The indicator is showing that the new price extreme was reached with a weaker difference between the fast and slow averages. That can matter, but it does not tell me when price will turn.
I want a price event after the divergence: a failed breakout, a close through a swing level or a break in the existing sequence of highs and lows. Without that, divergence is an observation rather than an entry rule.
MACD settings: 12, 26 and 9
The default settings are a useful starting point because most chart examples and platform defaults use them.
| Setting | Default | What changes when shortened |
|---|---|---|
| Fast EMA | 12 | MACD reacts more quickly to recent price changes. |
| Slow EMA | 26 | The comparison window becomes shorter. |
| Signal line | 9 | Crossovers appear sooner and more often. |
The numbers refer to candles, not fixed units of time. On a daily chart, 12 means 12 daily closes. On a five-minute chart, it means 12 five-minute closes.
Faster settings do not make MACD more accurate. They make it more sensitive. That can help when a strategy needs earlier information, but it also creates more crossovers in ordinary price noise.
I start with 12, 26 and 9. I change them only when I can test the same entry, invalidation and exit rules across enough trades, including spreads and slippage.
MACD formula
MACD line = 12-period EMA - 26-period EMA
Signal line = 9-period EMA of the MACD line
Histogram = MACD line - Signal lineMACD is not normalized. A reading of 20 can be small for one asset and impossible for another. Raw MACD values should not be used to compare momentum across instruments with different price scales.
My checklist before using MACD
- Price structure: Is the market trending or ranging?
- Zero line: Which side is MACD on, and is it staying there?
- Signal line: Is this a fresh momentum shift or another crossover near zero?
- Histogram: Is the gap expanding or contracting?
- Price trigger: What happened on the chart besides the indicator signal?
- Invalidation: Which price level proves the idea wrong?
- Costs: Is the expected move large enough to absorb spread, fees and slippage?
MACD is most useful to me when it compresses information I can already verify on price. If the indicator and the chart tell completely different stories, I do not let the indicator overrule the chart.
