RSI, MACD and Bollinger Bands - what they measure, why they lag, how they mislead

Three indicators appear on almost every trading screen in India, and almost everyone reaches for them before they understand them: RSI, MACD, and Bollinger Bands. Each has a number attached — a line crossing 70, a "crossover", a price tagging a band — and each of those numbers feels like an instruction. RSI at 78 seems to say sell. A MACD crossover seems to say buy. A band touch seems to say turn.

So the question for this module is precise. Each of these three tools measures a real thing. What, exactly, does each one measure — and does that measurement ever become a command? Or is the "signal" just the same price you were already watching, run through a named formula and re-drawn in a shape that feels authoritative?

The previous module established the family trait: an indicator is a formula on past price or volume, so it can never hold more information than its inputs. This module takes the three most famous members of that family and names them exactly — including the default settings the audit found people quoting without knowing.

Here is the whole thesis in one line, and everything below defends it: a threshold is not a command, and being "stretched" is not the same as being weak. RSI crossing 70 does not order a sale. A MACD crossover does not predict the next move; it re-describes momentum already sitting in the price. A Bollinger band touch does not signal a reversal; it reports that price is far from its own recent average. Each tool measures something genuine — momentum, trend relationship, dispersion — and each is routinely misread as an oracle that issues orders.

This matters because these three are where beginners most often hand their judgement to a line. The safest way to use any of them is to first ask how it would fool you — and the answer is always the same: by feeling like more than the price re-drawn.

Name the anatomy plainly and the mystique drains out. Each tool has a standard setting, an input, and a single thing it measures. illustrative

RSI — the default is 14-period. RSI (the relative strength index) looks at the last 14 closes and compares the size of the up-moves to the size of the down-moves. It maps that comparison onto a 0-to-100 scale. A high reading means up-moves have dominated recently; a low reading means down-moves have. The famous lines — above 70 "overbought", below 30 "oversold" — are conventions a designer chose, not thresholds the market respects. As an oscillator averaging the past, RSI lags, and in a strong trend it can sit above 70 for weeks.

MACD — the default is (12, 26, 9). MACD (moving average convergence divergence) is the gap between a fast 12-period and a slow 26-period EMA of the price. When the fast average pulls above the slow one, the gap is positive and momentum is up; when it falls below, the gap is negative. A third line — the 9-period signal line, an EMA of the MACD gap itself — is laid on top, and a "crossover" is simply the moment the gap crosses that smoothed version of itself. Every part is built from prices that already happened, so a crossover re-describes momentum already visible in the price. It arrives late by construction.

Bollinger Bands — the default is (20, 2 standard deviations). This is an overlay: a 20-day moving average with a band drawn two standard deviations above and two below it. Because the bands are set by recent volatility, they widen when price gets jumpy and narrow when it calms. Price tagging a band means "stretched versus its own recent range" — a measure of distance, not a reversal signal. In a real trend, price can ride the upper band while the band simply widens to follow.

Notice what the three share, because it matters more than what separates them: all three are averages of the past. RSI averages recent gains and losses; MACD subtracts one average from another and smooths the result again; Bollinger Bands average the price and measure how far it has strayed. Different questions, one raw material — the price you were already looking at.

None of this needs hard arithmetic — only the honest admission that averaging the past always arrives after the past. Lag is not a bug in these tools; it is their nature.

Take RSI. On the day a stock bottoms and starts to bounce, there are no rising days yet inside its 14-day window — so RSI has nothing to lift it. It can only turn up once several rising days have accumulated. It follows the low; it never leads it.

Take MACD. It subtracts a 26-period EMA from a 12-period EMA — two lagging averages — and then smooths that gap again with a 9-period signal line. A crossover is therefore an average of an average of an average, crossing another smoothed line. By the time it fires, the momentum it "signals" has already been in the price for days. That is why a crossover so often feels like it arrived just after the move you can already see on the candles.

Take Bollinger Bands. The middle line is a 20-day average, so it turns after the price turns, and the bands — set by the standard deviation of those same 20 days — only widen once volatility has already risen. They confirm a change in behaviour; they cannot warn of one.

Here is the central claim, made touchable. Below is a composite price with a 14-period RSI panel underneath it. illustrative Switch between a strong uptrend and a sideways range, and watch the same 70 line mean two completely different things.

In the uptrend, RSI crosses 70 early and then stays there — for many days — while the price keeps climbing. If you had obeyed "70 means sell," you would have thrown away most of the move. The RSI is not wrong; it correctly reports that up-days have dominated. What is wrong is reading a description of strength as a warning of weakness.

In the range, the very same 70 and 30 lines flip on and off again and again, firing signal after signal, none of which starts a lasting move. The threshold did not change. The market state did.

The lesson the widget makes on its own: overbought is not automatic weakness, and oversold is not automatic strength. These are readings of the recent past. Whether they matter depends entirely on the state of the market you are in — and the indicator never tells you which state that is. You have to.

Because all three tools measure the recent past, the same reading inverts depending on the market state around it. This is the heart of why they mislead: the number stays put while its meaning flips.

An RSI of 75 is a shrug in a strong uptrend — the stock is strong, so momentum is strong — and a genuine caution in a tired, sideways market where such a reading has repeatedly marked local tops. A MACD crossover confirms a move worth riding in a trending market and whipsaws you to death in a choppy one, firing and reversing every few days. A Bollinger band touch is normal, even expected, when price is trending and riding the band, but is a real stretch signal when price is oscillating inside a stable range.

So "what did the indicator say?" is never the whole question. The whole question is: what does this tool measure, what does it lag, and what state is the market in right now? The reading is the same in both columns. Only the context tells you whether it is a shrug or a siren.

Take one concrete case, because it shows how a MACD crossover gets misread as either prophecy or failure. illustrative

A composite stock has fallen, based, and begun to recover. Six days into the bounce, the MACD (12, 26, 9) finally posts a bullish crossover — the gap between the fast and slow EMA crosses above its 9-period signal line. One beginner sees the crossover and thinks: there — MACD called the turn. Another, who had watched the price base and bounce directly, thinks: useless — it crossed six days after the low.

Both misread the same event. The crossover did not call the turn; it is the moment two lagging averages, plus a smoothing of their gap, finally agreed that momentum had shifted — which could only happen after enough rising days had accumulated. It is not prophecy (it followed the price) and it is not failure (following the price a few days late is exactly what a triple-smoothed line does). It is a lagging summary behaving normally.

Now imagine you sped MACD up to catch the turn sooner. It would have crossed earlier here — and it would also have fired two false crossovers during the choppy base the week before, each one snapping back. The slower default missed those false alarms at the cost of a later real signal. Faster and cleaner are trade-offs, not moral qualities, and no setting gives you both.

Being clear about what each tool measures fences off the errors that cost beginners the most.

None of the three can create information that is not in the price. RSI, MACD, and Bollinger Bands are all built from past closes — so they can reflect momentum, trend relationship, and dispersion, but never the earnings, the debt, the promoter's plans, or the next big buyer's decision. They re-arrange the visible past; they cannot import the invisible future.

None of them can lead the price they are built from. A crossover, a threshold, a band touch — all are arithmetically downstream of the prices that produced them. They can confirm a move in progress and summarise a trend. They cannot get ahead of the data that defines them.

None of them issues a command. "Overbought," "crossover," "band touch" are re-descriptions of what price already did, not instructions about what you should do. Used honestly, they are tools for timing an action you already had good grounds to take — the subject of a later module — never the reason for the action itself.

And no combination of them rescues a weak thesis. Stacking RSI, MACD, and Bollinger Bands is not three witnesses agreeing; it is one dataset — the price — read three ways. Their "agreement" is largely built in.

The same handful of misreadings catch users of these three again and again. Name them and they lose their grip.

1. Treating a threshold as a command. 70/30 on RSI are round numbers a designer picked, not laws. A strong stock can stay "overbought" for weeks. The line describes momentum; it does not order a trade.

2. Reading a crossover as a prediction. A MACD crossover is a smoothing of a gap between two lagging averages. It re-describes momentum already in the price. It confirms, late; it never forecasts.

3. Reading a band touch as a reversal. Tagging a Bollinger band means "stretched versus the recent average," not "about to turn." In a trend, price rides the band while the band widens to follow.

4. Trusting divergence as a timer. A divergence — momentum easing while price makes new highs — is a caution flag, not a countdown. It can persist for weeks, or resolve with no reversal at all.

5. Mistaking wider bands for a breakout. Band width measures volatility, full stop. Volatility rises in real breakouts and in violent fake-outs alike. Width is dispersion, not direction.

6. Stacking correlated tools and calling it confirmation. RSI, MACD, and Bollinger Bands are all built from the same price. Their agreement is mostly guaranteed by construction, not independent evidence.

7. Fitting settings until one agrees. Cycle through 14, 9, 21 on RSI, or a dozen MACD sets, and one will eventually flatter a decision you already made. That is shopping for a signal, not receiving one.