Ask ten technical traders which indicator sits on their chart beneath price, and a majority will name the same one: MACD, short for Moving Average Convergence Divergence. Invented by Gerald Appel in the late 1970s, it has survived five decades of market evolution because it answers a question every trend follower cares about — is momentum building or fading? — using nothing more exotic than two moving averages and a little subtraction.
This guide walks through what MACD is, how each of its three components is built, what the classic signals mean, where the indicator shines, and — just as important — where it routinely misleads people.
What MACD actually measures
A moving average smooths price so you can see the trend. A pair of moving averages — one fast, one slow — lets you see the trend on two clocks at once. When the fast average pulls away from the slow one, the move is accelerating. When the gap narrows, the move is losing steam.
MACD simply measures that gap and plots it as a line. Everything else — the signal line, the histogram, the crossovers — is machinery built on top of this single idea: the distance between a fast and a slow exponential moving average is a readout of momentum.
Because it is built from moving averages, MACD belongs to the family of trend-following momentum indicators. It is not an oscillator bounded between 0 and 100 like RSI; it can drift as far above or below zero as price momentum carries it. That unbounded quality is a strength (it never artificially caps a strong trend) and a weakness (you cannot compare raw MACD values across stocks that trade at different prices).
The three components
A standard MACD panel shows three things, and each answers a different question.
1. The MACD line
The MACD line is the fast EMA minus the slow EMA — in the standard configuration, the 12-period exponential moving average minus the 26-period one. When the 12 EMA sits above the 26 EMA, MACD is positive; when below, negative. The further apart they are, the larger the absolute value.
- MACD above zero — the shorter-term average is above the longer-term one; the recent trend is up.
- MACD below zero — the reverse; the recent trend is down.
- MACD rising — the fast average is pulling away upward, i.e. upside momentum is building, even if the value is still negative.
2. The signal line
The signal line is a 9-period EMA of the MACD line itself. It smooths the smoothed. Its job is purely mechanical: to give the MACD line something to cross. When the MACD line crosses above its signal line, short-term momentum has turned up relative to its own recent history; when it crosses below, momentum has turned down.
3. The histogram
The histogram plots the difference between the MACD line and the signal line as vertical bars around a zero axis. It is the most sensitive of the three components — the derivative of a derivative, in effect — and it is where changes show up first. A shrinking positive histogram says an uptrend in momentum is decelerating before any crossover happens. Many experienced users watch the histogram's direction more closely than the lines themselves.
Walking through the calculation
Suppose a stock closes at 500 and has been drifting up for weeks. Its 12-day EMA might sit at 496 and its 26-day EMA at 488.
1. MACD line = 496 − 488 = +8. 2. Suppose the 9-day EMA of the MACD line's recent values is +6.5 — that is the signal line. 3. Histogram = 8 − 6.5 = +1.5, drawn as a bar above zero.
Now imagine the rally stalls. The fast EMA flattens quicker than the slow one, so the gap narrows: MACD slips from +8 to +6 while the signal line, being a lagging average of MACD, is still around +7. The histogram flips negative (−1) even though the stock is still above both moving averages and MACD is still positive. That early flip is the histogram doing its job: momentum is fading at the margin.
Note what this also demonstrates: every layer is a rate of change of the one beneath it. Price → EMAs → their gap (MACD) → the gap's own average (signal) → the gap between those (histogram). Each layer reacts faster but is noisier.
The classic signals
Signal-line crossovers
The best-known MACD event: the MACD line crossing its signal line. A bullish crossover (MACD crossing above the signal) says short-term momentum has turned up; a bearish crossover says the opposite. On end-of-day charts of trending large caps this is a serviceable description of momentum turns. In choppy, sideways markets it fires constantly and unprofitably — this is the single most important caveat with MACD, and we return to it below.
Zero-line crossovers
When the MACD line itself crosses zero, the 12 EMA has crossed the 26 EMA. This is a slower, more deliberate event than a signal-line cross — closer in spirit to the 50/200-DMA relationship, though on a much faster clock. Some trend followers treat the zero line as a regime filter: only take note of bullish signal crosses when MACD is above zero (trend and momentum agreeing), and vice versa.
Histogram reversals
The histogram peaks before the MACD line does. A sequence of tall positive bars followed by visibly shorter positive bars means the advance is decelerating; the crossover, if it comes, arrives later. Traders who act on histogram contraction accept more noise in exchange for earlier information — a classic speed-versus-reliability trade-off that has no free lunch.
Divergence
Divergence is when price and the indicator disagree. A bearish divergence: price prints a higher high, but MACD prints a lower high — the second push up had less force behind it. A bullish divergence is the mirror image at lows. Divergences are best treated as context, not triggers: they warn that a trend is tiring, but tired trends can keep grinding for a long time. In strong Indian bull runs, bearish MACD divergences have persisted for months while the index kept climbing.
Choosing settings — and why the defaults endure
The canonical 12/26/9 settings date from an era of six-day trading weeks (12 ≈ two weeks, 26 ≈ a month). Nobody has repealed them because MACD's usefulness is structural, not numerical. Still, it helps to know the dials:
- Shorter settings (e.g. 5/35/5 or 8/17/9) — faster, better for swing traders, more whipsaw.
- Longer settings — slower and smoother, fewer but later signals, better for positional views.
- Weekly charts with default settings — a favourite of longer-term trend followers; a weekly MACD cross is a meaningfully big event.
Changing settings does not change MACD's fundamental character: it will always lag price (it is built from lagging averages) and always chop in sideways markets.
Where MACD works — and where it fails
MACD is a trend tool. Its signals have historically been most informative when:
- the stock or index is actually trending (up or down) on the timeframe you trade;
- the signal agrees with a larger context — e.g. a daily bullish cross while the weekly MACD is already positive;
- the crossover happens away from major overhead resistance or support, leaving room for the move to develop.
It fails, predictably and repeatedly, when:
- the market is range-bound. The averages braid around each other and the crossovers become coin flips with commission costs.
- you read raw values across stocks. A MACD of +8 on a ₹2,500 stock is not "stronger" than +2 on a ₹300 stock; the scale depends on price. Compare shapes, not levels — or normalise (some platforms offer a percentage variant, the PPO, for exactly this reason).
- you expect it to call tops and bottoms. MACD is late by construction. It confirms turns; it does not anticipate them. Anyone promising otherwise is selling something.
MACD versus RSI — rivals or teammates?
The two most popular momentum tools answer different questions. RSI measures the internal balance of recent up-days versus down-days on a bounded 0–100 scale — good for spotting stretched conditions. MACD measures the relationship between two trend proxies on an unbounded scale — good for tracking the life cycle of a trend. They disagree often, and usefully: a stock can be short-term overbought on RSI in the early innings of a trend that MACD says is just getting going. Many workflows pair them: MACD (or the 200-DMA) to define the regime, RSI to time entries within it.
Using MACD in a screener
On an end-of-day platform like TaurEye, MACD becomes a filtering tool rather than a watching tool:
- MACD histogram above zero — surfaces stocks where momentum currently favours the upside.
- MACD bullish cross today — the histogram crossed above zero in the latest session; a fresh momentum turn to investigate. The Screener exposes this as a one-click signal filter.
- Combine with trend filters — e.g. bullish MACD cross AND price above the 200-DMA AND relative volume above 1.5×. Stacking conditions this way keeps you out of counter-trend noise, at the cost of fewer candidates.
The output of any such screen is a shortlist for research, not a buy list. Two stocks with identical MACD events can have opposite futures depending on earnings, sector flows and the broader market.
A worked reading routine
A practical end-of-day routine using MACD might look like this:
1. Establish the regime. Is the index above its 200-DMA? Is the weekly MACD positive? If not, treat bullish daily signals with suspicion. 2. Screen. Run a bullish-cross screen with liquidity and trend filters attached. 3. Read each chart. Where did the cross occur — after a long decline (potentially early), mid-range (noise-prone), or on a pullback within an uptrend (the textbook case)? 4. Check the histogram's story. Was momentum contracting for weeks before the cross, or did it flip on one volatile session? 5. Decide risk first. Where is the level that proves the idea wrong? If that level is far away, the signal may not be actionable at your risk tolerance regardless of how pretty the crossover looks.
Common mistakes
- Trading every crossover. In a sideways market this is a machine for converting capital into brokerage.
- Ignoring the timeframe hierarchy. A bullish daily cross inside a bearish weekly trend is a lower-probability event than the same cross with the weekly wind at its back.
- Comparing MACD values between stocks. Unbounded indicator, price-dependent scale.
- Treating divergence as a timing signal. It is a warning light, not a brake pedal.
- Forgetting that MACD lags. By the time a crossover prints, part of the move has already happened. That is the price of smoothing.
MACD through a full market cycle
Indicators behave differently in different regimes, and it is worth rehearsing MACD's personality in each phase of a cycle before trusting it with decisions.
Early bull phase. Coming out of a long decline, the first bullish zero-line cross on the weekly chart is often the single most useful MACD event in the entire cycle — it marks the point where the medium-term average finally overtakes the long-term one after months of repair. Daily signals during this phase tend to be productive because pullbacks are shallow and momentum keeps re-asserting.
Mature bull phase. The trend is established, everyone can see it, and MACD spends most of its time positive. Signal-line crosses become pullback markers rather than trend calls: MACD dips toward its signal line as the stock digests gains, then re-crosses upward as the trend resumes. Bearish crossovers here are frequently head-fakes — this is where mechanically shorting every bearish cross gets expensive.
Distribution and topping. The tell is repetition: price grinds to marginal new highs while each MACD peak comes in lower than the last — the multi-month bearish divergence. No single divergence is decisive, but a sequence of them, especially with weekly momentum flattening, describes a trend running on fewer engines.
Bear phase. Everything inverts. MACD lives below zero, bullish daily crosses become counter-trend bounces that fade near the zero line, and the zero line itself acts like a ceiling. Traders who flip their playbook (fade strength rather than buy it) find MACD as useful on the way down as on the way up — but most retail participants do not flip, and the indicator gets blamed for what is really a regime-reading failure.
Sideways churn. The blunt answer: MACD is close to useless in a genuine range, and the sooner you recognise a range, the sooner you can either stand aside or switch to range tools (support/resistance, %B). The bands of profitability for MACD systems in backtests almost always come from trending segments; the losses come from the chop.
Common questions
Does MACD work on intraday charts? The arithmetic works on any timeframe, but noise scales up as the timeframe shrinks. On a 5-minute chart the indicator responds to microstructure — order flow, news blips — where whipsaw dominates. TaurEye is an end-of-day platform for this very reason: daily and weekly signals carry more signal per unit of noise for non-professional traders.
Is a bigger histogram bar better? Bigger bars mean faster momentum, but "better" depends on position in the move. Explosive histogram expansion at the start of a trend (after a squeeze or base) is constructive; the same expansion after months of rally is often a climax signature.
Why did the crossover print but the stock fell anyway? Because MACD is a description, not a cause. Roughly half of all daily bullish crossovers in a flat market resolve lower. The indicator's value comes from conditioning — pairing it with trend, liquidity and level context — not from the event alone.
Should I use MACD or the 50/200-DMA cross? They are the same idea on different clocks. The golden/death cross uses simple averages and long windows (a regime tool that changes a handful of times a decade for an index); MACD uses exponential averages and short windows (a tactical tool that changes monthly). Many workflows use both: the slow cross to define the campaign, MACD to time engagements within it.
Can MACD be used for exits? Yes — arguably better than for entries. A trailing exit on a bearish signal-line cross gives a trend room while defining when momentum has objectively rolled over. It will always give back some open profit (lag again), which is the tuition every trend-following exit pays.
Where this leaves you
MACD endures because it compresses a genuinely useful idea — the convergence and divergence of two trend proxies — into one glanceable panel. Used as a regime and momentum descriptor, stacked with trend and liquidity filters, and read with an awareness of its lag, it earns its place on the chart. Used as an oracle, it disappoints exactly as often as any other single indicator.
This article is for learning, not advice. An indicator can only describe what prices have already done — nothing here predicts what they will do next. Please do your own research, and speak to a SEBI-registered adviser before putting money at risk.