How beginners should read MACD (and why not to predict with it)

MACD is probably the indicator beginners have heard the most about — and the one most often turned into a myth. Short videos love a line like "MACD bullish cross, go all in", which makes it sound like picking up free money. This note pulls it apart, then tells you honestly what it can and can't do: it's a useful mirror, but it is nowhere near a crystal ball for which way prices go.

The three questions beginners ask most, with the short answers first:

What MACD is really measuring

Keep one sentence in your head first: MACD measures the "momentum" of a move — whether the push behind it is strengthening or fading. Two coins can both be going up, but is one drifting up slowly while the other is ripping higher? That difference in oomph is what MACD is trying to show you.

How does it get there? It's pretty plain once you see it: take a fast moving average, subtract a slow one, and you've got the gap between them. When price accelerates higher, the fast line pulls away from the slow line, the gap widens, and that says upward momentum is building. When the climb cools off, the two lines drift back together, the gap narrows, and momentum is fading. So — if you've read how to read moving averages, you'll spot that MACD is basically a spin-off of moving averages, not some brand-new beast. It just takes the distance between two averages and draws it out as its own picture.

The three parts: DIF, DEA, the histogram

Open MACD and you'll see two lines and a row of little bars in the window below the price. Don't let the jargon scare you — the mapping is simple:

Of these three, the histogram is the one a beginner should watch most, because it's the most direct: bars above the zero line lean toward upside momentum, bars below it lean toward downside. More to the point, watch whether the bars are growing or shrinking — longer and longer means the push in that direction is building; bars starting to shrink mean the push is fading, and even if price is still rising, the steam behind it may already be running out.

For the DIF and DEA lines, a rough impression is enough for now: DIF is twitchy and hugs the action, DEA is sluggish and trails behind, and where they cross is the crossover we'll talk about next.

Crossovers and "divergence"

MACD crossovers work on the same idea as the moving-average ones, just with different leads:

• Bullish cross: DIF (the fast line) crosses up through DEA (the slow line), and the histogram flips from negative to positive. Often read as a hint that momentum is turning up.
• Bearish cross: DIF crosses down through DEA, and the histogram flips from positive to negative. Often read as a hint that momentum is turning down.

Then there's a word that sounds mystical — "divergence" — but it's not hard. Price makes a new high, but MACD doesn't follow it up and instead rolls over: that's bearish divergence, and the idea is "price is still pushing up, but the strength behind it isn't keeping pace." It's often read as a sign the climb is running low on fuel. Flip it — price makes a new low while MACD doesn't — and you've got bullish divergence.

Notice I keep saying "often read as." That's on purpose. Divergence is only a hint, not a promise that a reversal is about to land. Price carrying on in the same direction after divergence shows up is extremely common.

Read crossovers and divergence as descriptions of momentum, not as buy/sell orders. MACD is telling you "the push has changed." Where the price goes next, it cannot tell you.

Why you can't predict with it

Let me be blunt: MACD can't be used to predict prices. The reason is the same as with moving averages — because it's built out of moving averages, it inherits the exact same flaws:

• It lags. MACD is calculated from past prices, so it runs behind the price. By the time a bullish cross is confirmed, price has usually already climbed a stretch; by the time a bearish cross appears, it's often already dropped. Trading in and out off the signal makes it easy to buy high and sell low.
• It breaks down in a sideways market. When price grinds flat, DIF and DEA tangle together and cross back and forth, and the crossovers come thick and fast — all of them fakes. Trade them and you get sawn off both ways.

At the end of the day, MACD is a rear-view mirror — it can tell you what just happened, but it's helpless about the road straight ahead. No indicator on earth reliably predicts the future, and if MACD really could call which way prices go, the person who invented it would have quietly gotten rich rather than writing it into a textbook. Reading candlestick charts for beginners makes the same point over and over: reading a chart means reading a record of what already happened, not fortune-telling.

How a beginner should actually use it

So the flaws are out in the open — that's not a reason to throw it away, it's a reason to put it in the right place. Three suggestions for a beginner:

1. Use it as a sidekick, not the lead. First read price and the moving averages to judge the big picture (uptrend / downtrend / sideways), then use MACD to feel "how strong is the momentum right now." It's a supporting role.
2. Watch the bars changing, not just the cross. Bars shrinking means the push in the current direction is fading — that's often earlier and more useful than "which day did it cross."
3. Don't trust it in a sideways market. When price is chopping inside a range, just ignore its crossovers — they're basically noise.

And the most important one: never make a decision on a single indicator. MACD, moving averages, RSI — each is just one angle on the same stretch of price. What actually decides whether you win or lose isn't whether some indicator is "accurate," it's whether you've managed your risk: before you buy, be clear on where you'd admit you're wrong and how much you can lose at most — see the risk calculator. Want to add one more angle? Read RSI and "overbought / oversold" next.

Finding MACD on the OKX screen

Take OKX (formerly OKEx) as the example — both the web and the app can bring up MACD, and it comes with a demo account, which makes it good for watching over and over:

Leave the settings on default — don't go chasing "the optimal parameters," because that thing doesn't exist. Get comfortable reading the histogram, and your grasp of MACD is already ahead of most people who only know how to shout "bullish cross, all in."