Indicators — Lessons

Chapter 2: Stochastic RSI — The Second Derivative

Chapter 1 ended on a warning: RSI can sit in overbought or oversold territory for weeks without ever really moving. Stochastic RSI exists because two analysts got tired of exactly that problem, so they built an indicator of an indicator to fix it.

Streamline View: How StochRSI Is Built and Read

Stochastic RSI (StochRSI) asks where today’s RSI sits between its own recent low and high, instead of where price sits. The rule that matters most is that it is a magnifier: it reaches its extremes far more often than RSI, so a touch of 80 or 20 is common and a lone touch carries little weight. The main way it goes wrong is trusting that an extreme means the move is over, when in a strong trend it can stay pinned at 0 or 100 for many bars.

Second outcome. Suppose RSI has been tightly held between 48 and 52. At 50 the raw reading is 50. If RSI then edges up one point to 51, the reading jumps to 75, a move of 25 points from a one-point change. That is the twitchiness at work, and it is why a StochRSI above 80 after a quiet stretch says little about the market itself.

In plain words: StochRSI shows where RSI stands inside its own recent range, it fires often by design, and the useful signal is a held %K and %D crossover in an extreme zone, read with the trend. Education, not advice.

History of StochRSI: Two Older Ideas Joined in 1994

StochRSI joins two older ideas: George Lane’s stochastic formula for where a close sits in a range, and Wilder’s RSI. Its own history is short, a 1994 book and then a long run of platform defaults, but the two parents explain both its speed and its habit of crying wolf. Nothing in the record suggests the combination was meant to stand alone; the authors framed it as a way to see RSI more sharply, and later platforms turned it into a default panel.

• The 80 and 20 lines were a design choice. Chande and Kroll widened RSI’s 70 and 30 on purpose, to get more signals. So an overbought reading here is a routine event, and the original authors expected it.
• The 0 to 100 scale and the smoothing came later. The 1994 form had no %K and %D smoothing and ran from 0 to 1. Platforms multiplied it by 100 and added the lines borrowed from Lane’s stochastic, so two charts can look different with the same name.
• It stacks two lags. RSI is already an average, and StochRSI measures a range of that average. A sharp price turn reaches the final line only after passing through both layers, even though the line itself looks twitchy. That is why a held crossover, not a single touch, is the sturdier thing to watch, and why the reading is best set beside the trend instead of used alone.

Who Built This, and Why

Stochastic RSI was developed by Tushar Chande and Stanley Kroll and published in their 1994 book The New Technical Trader. Their stated reasoning was blunt: regular RSI can oscillate between 80 and 20 for extended periods without ever reaching a true extreme, which makes it slow to flag genuine overbought or oversold conditions. Their fix was to stop applying the Stochastic formula to price, like the original Stochastic Oscillator does, and instead apply it to RSI's own output.

The Formula: An Indicator of an Indicator

StochRSI takes the last several days of RSI readings (not price, RSI itself) and asks one question: relative to its own recent high and low, where does today's RSI sit right now?

Look at what that number is actually saying. A raw RSI of 50 looks perfectly neutral. Textbook middle-of-the-road. But StochRSI of 30 tells you RSI itself has fallen a long way from its own recent peak of 78, down toward the bottom third of its own range. That's information the raw RSI number alone simply doesn't carry.

Default Settings Almost Everyone Uses

The near-universal default across TradingView, most brokers, and every major charting platform is 14, 14, 3, 3: a 14-period RSI, fed into a 14-period Stochastic lookback, with the raw result smoothed into a %K line using a 3-period moving average, and %D as a further 3-period moving average of %K, the exact same %K/%D relationship the original Stochastic Oscillator uses, just built on RSI instead of price.

Why It's So Much Twitchier Than RSI

This is the entire point of the indicator, and also its biggest risk. Because StochRSI measures RSI relative to its own recent range instead of on a fixed 0–100 scale, it reaches its own overbought and oversold extremes far more often. Charting-platform technical documentation describes it as "dramatically" increasing signal count compared to plain RSI. Take Chapter 1's real example: the S&P 500's 14-day RSI went from oversold to overbought over 12 trading days following the March 30, 2026 low. StochRSI, watching that same real climb, would have registered its own overbought reading well before RSI itself ever touched 70, and likely printed more than one %K/%D crossover along the way up, some of them well before the move was actually over.

The Real Cost of Over-Trusting It

A widely shared trading-education story tells of a trader buying every StochRSI reading below 20 during what turned out to be a strong, sustained downtrend, averaging down five separate times into the close as the stock kept falling anyway. The writer calls it a "$4,250 lesson." It's a useful, honest illustration of the core risk: StochRSI hitting an extreme doesn't mean the extreme is over. In a strong trend, it can hit 0 or 100 and simply stay pinned there.

How Professionals Actually Use It

StochRSI is almost never used as a standalone trigger. Most traders treat a %K/%D crossover in the extreme zones as the actual signal, not a single bar touching 80 or 20; a crossover confirms the move has actually turned, where a bare touch can reverse instantly. It's also read alongside the underlying trend, just like RSI itself. A StochRSI dip to 0 during a strong uptrend is read very differently than the same dip during a downtrend. Many traders use it as a fast-moving early warning layered on top of regular RSI, not as a replacement for it.

Common Mistakes

• Trading every single touch of 80 or 20 as an automatic signal: in a strong trend, StochRSI can stay pinned at an extreme for many bars in a row.
• Averaging into a losing position because "it's oversold," the same real mistake behind the $4,250 lesson above.
• Reading StochRSI in isolation instead of alongside regular RSI or the broader trend — it was built to add sensitivity, not to replace the tools that give it context.
• Treating a single bar's reading as a signal instead of waiting for the %K/%D crossover, which filters out a meaningful share of the noise.

Watching This for Free

StochRSI ships as a free, built-in study on TradingView's free plan and most retail broker charting packages, almost always defaulted to the same 14, 14, 3, 3 settings covered above, no different in cost or access from RSI itself.

1. 1Add StochRSI with the default 14, 14, 3, 3 settings alongside regular RSI on the same chart.
2. 2Treat a %K/%D crossover in the extreme zones as the real signal, not a single bar simply touching 80 or 20.
3. 3Check the broader trend before acting: an extreme StochRSI reading against the trend is far less reliable than one that agrees with it.
4. 4Never average into a position purely because StochRSI is pinned at an extreme; pinned can mean "about to turn" or it can mean "strong trend, stay away."
5. 5Use it as an early, noisier layer on top of RSI's slower, steadier read, not as a full replacement for it.

Try It Yourself: Turn an RSI Series Into a StochRSI Reading

The fastest way to really understand StochRSI is to build one from a RSI history you already have, once, by hand.

1. 1Pull up any chart with RSI already on it and write down the last 14 days of RSI values (you can read these straight off the RSI line, no recalculation needed).
2. 2Find the highest and lowest RSI values in that 14-day window.
3. 3Take today's RSI, subtract the lowest value, and divide by the highest value minus the lowest value.
4. 4Multiply by 100: that's the raw StochRSI reading before smoothing.
5. 5Add the platform's built-in StochRSI to the same chart and check your hand calculation against its raw (unsmoothed) line. They should land in the same neighborhood.

That's the real skill this chapter is teaching: StochRSI isn't a better version of RSI, it's a magnifying glass held up to RSI, useful for catching turns early, dangerous if you forget that a magnifying glass also magnifies every false alarm along with the real ones.

Visualize It: Where Is RSI Inside Its Own Range?

StochRSI asks where today’s RSI sits between its recent low and high. Set the three values.

Two Analysts Who Thought RSI Wasn't Sensitive Enough

Stochastic RSI was built by Tushar Chande, an engineer turned quantitative analyst, and Stanley Kroll, a veteran commodities trader who'd already written several trading books of his own. Their complaint, laid out in their 1994 book "The New Technical Trader," was specific: regular RSI could sit between 20 and 80 for weeks without ever flagging a real extreme. Their fix was procedural rather than clever: run the Stochastic formula on RSI's own output instead of on price. It's a small idea that became a standard oscillator in its own right.

Cypher Guide: StochRSI Rolled Across a Window, With Range Sensitivity

This guide fixes the vocabulary of StochRSI and then works the calculation across a rolling window: the raw reading, the two smoothing steps, what the window length does, and what happens when the RSI range collapses. The chapter’s 14-day RSI series is extended by six hypothetical days. All figures are hypothetical.

Mechanism one: a rolling window. The first raw value needs 14 RSI values, so it appears on day 14. The two smoothing steps each cost two more bars, so %D first exists on day 18. Each row below drops the oldest RSI value from the window and adds the newest.

Read across the row for day 16. RSI fell from 48 to 45, only three points, yet the raw value dropped from 16.7 to 0.0 because 45 was the lowest value in the window. A three-point RSI move produced a full-scale swing in the raw line. The smoothed lines soften this: %K at that day is 15.6, and %D does not exist yet at that point.

The same RSI of 59 reads 42.4 on a 14-bar window and 87.5 on a 5-bar window. A shorter window cannot see the earlier low or high, so the reading depends on how much history the window holds. Quote the window with every value.

Mechanism two: why a small range magnifies. The raw value moves by 100 ÷ range points for each point of RSI. The table holds the RSI move at 2 points and changes only the range.

Read across the row for range 5. A quiet RSI makes StochRSI the most violent when nothing much is happening, which is the mechanical reason it is twitchier than RSI and why a flat RSI can still print 0 and 100 in the same week.

In plain words: StochRSI re-scales RSI by its own recent high and low, so the same RSI can read very differently depending on the window and the range. Education, not advice.

Closing Rundown: What StochRSI Adds to RSI and What It Costs

Chapter 2 set out to show how StochRSI is built on top of RSI and why it behaves so differently from it. The one idea is that it asks where today’s RSI sits inside its own recent range, which makes it a magnifier of RSI and not a better version of it.

• A range inside a range. StochRSI = (RSI − lowest RSI) ÷ (highest RSI − lowest RSI) × 100. With a window from 38 to 78 and an RSI of 50, the raw reading is 30.0, even though 50 looks neutral on RSI itself.
• Small ranges magnify. Each point of RSI moves the raw line by 100 ÷ range points. A 2-point RSI move is worth 5 raw points on a range of 40 and 40 raw points on a range of 5.
• Extremes are routine. The 80 and 20 lines were chosen on purpose to produce more signals, so a lone touch carries little weight. In a strong trend the line can stay pinned at 0 or 100 for many bars.
• The crossover is the signal. A %K and %D crossover in an extreme zone that holds is sturdier than a bar that only touches 80 or 20. The usual settings are 14, 14, 3, 3.
• Quote the window. The same RSI of 59 reads 42.4 on a 14-bar window and 87.5 on a 5-bar window. Smoothing choices also differ between platforms.
• Averaging down is the costly habit. The $4,250 story showed five purchases at readings under 20 in a downtrend. An extreme reading does not mean the extreme is over.

Chapter 3: MACD — Moving Average Convergence Divergence

RSI and StochRSI are both boxed into a fixed 0–100 scale. MACD isn't — it has no ceiling and no floor, because it isn't measuring where price sits on a scale at all. It's measuring the gap between two moving averages, and how that gap is changing. That's a genuinely different tool, not just another flavor of the same one.

Streamline View: How MACD Is Built and Read

MACD, Moving Average Convergence Divergence, is the gap between a fast and a slow exponential moving average, plotted with a smoothed copy of itself and the bars between the two. The rule that matters most is that it measures how fast the averages are pulling apart or together, not where price sits on a scale. The main ways it goes wrong are looking for overbought and oversold levels that do not exist on it, and trading every crossover in a market that is going sideways.

Second outcome. To compare two markets, divide the MACD line by the slow average. Here 3.75 ÷ 178.65 is 2.1%. The same MACD of 3.75 on a stock whose 26-day EMA is 20.00 would be 18.8%, a far stronger gap. The raw number alone cannot say that.

In plain words: MACD plots how far apart two averages are and how fast that gap is changing, the histogram shows the change early, and no level on it means overbought or oversold. Education, not advice.

History of MACD: From Exponential Smoothing to the Histogram

MACD is a gap between two exponential averages, so its history has two layers: the smoothing method it rests on, and the trader who turned it into a timing tool and the later hands that changed how it is drawn. The bar chart most traders look at first today, the histogram, is the newest piece. Appel did not patent MACD or try to control it, according to the chapter, so it spread by being copied into charting software and became the shared default it is now.

• The histogram is an add-on, not the original. It was built to anticipate the signal-line cross, so it is an early warning of a cross that has not happened yet. Bars that shrink do not by themselves mean the lines will cross.
• The settings come from an older calendar. The chapter ties 12, 26 and 9 to a six-day trading week. The numbers stuck because everyone copied them, so using the same ones lets a chart match what other traders see. Other settings exist for short-term work, but they are a personal choice and make a chart harder to compare.
• A price-unit tool needs a percent version for comparisons. Appel’s MACD is in price units. A percentage version divides the gap by the slower average, and is preferred when comparing markets with very different prices. Without it, a raw reading of 3.75 cannot be compared across a $20 stock and a $2,000 one. The same limit explains why MACD has no overbought or oversold line: a price-unit gap has no fixed ceiling.

Who Built This, and Why

MACD was developed by Gerald Appel in the late 1970s, and later documented in his 2005 book Technical Analysis Power Tools for Active Investors. The now-standard histogram (the bar chart showing the gap between MACD and its own signal line) wasn't part of Appel's original version. Thomas Aspray added it in 1986, later publishing the idea (1988) in a trading magazine, specifically to make MACD's own turning points easier to see before the signal-line crossover actually happened.

The Formula: Two EMAs and the Gap Between Them

MACD starts with two exponential moving averages of price (a fast one and a slow one) and simply subtracts the slow one from the fast one. That gap is the MACD line. A second, smoothed average of that gap becomes the signal line. The difference between those two is the histogram.

The 12-day EMA sitting above the 26-day EMA is what "convergence/divergence" actually refers to. The two moving averages are diverging apart, and MACD is simply plotting the size of that gap as its own line.

Why 12, 26, and 9?

Those numbers look arbitrary, but they aren't. Appel built MACD back when trading weeks still ran six days. 12, 26, and 9 represented two weeks, one month, and a week and a half on that old six-day calendar. Markets have run five-day weeks for decades now, which means the "clean" modern equivalents would actually be slightly different numbers, but 12/26/9 became the standard everyone uses, so deviating from it mostly just makes your chart harder to compare with everyone else's.

Reading Crossovers

There are two different crossovers MACD gives you, and they mean different things. A signal-line crossover (MACD crossing above or below its own signal line) is the faster, more frequent signal, and it's exactly what the histogram flipping from negative to positive (or back) is showing you visually. A centerline (zero-line) crossover, MACD crossing above or below zero, is slower and heavier: it means the fast EMA has actually crossed the slow EMA outright, not just that momentum is shifting within an existing trend.

A Real, Very Recent Track Record

Bitcoin's daily MACD histogram gave a real, dated, and unusually clean track record of bearish crossovers over the winter of 2025 into 2026. On November 3, 2025, MACD turned bearish with Bitcoin trading around $106,000; by November 21 it had fallen to roughly $80,000. On January 20, 2026, MACD flashed bearish again, this time with Bitcoin near $90,000; by February 6 it had dropped to nearly $60,000. By late March 2026, the histogram had crossed below zero a third time.

The Big Limitation: No Overbought or Oversold

Unlike RSI or StochRSI, MACD has no fixed scale and no built-in overbought or oversold zone. There's no equivalent of "70" or "80" to watch for. A strong trend can push MACD to a reading nobody's ever seen on that chart before, and it still wouldn't technically be "overbought." That's also why MACD values can't be meaningfully compared across two different stocks: a MACD of 3.75 means something different on a $20 stock than it does on a $2,000 one, since it's built from raw price, not a normalized percentage.

How Professionals Actually Use It

MACD is most often used as a trend-confirmation tool sitting alongside a bounded oscillator like RSI or StochRSI, not as a replacement for either. The histogram's slope (whether the bars are growing or shrinking, even before a full crossover happens) is watched as an early read on fading momentum. And just like RSI, MACD divergence, price making a new high or low that MACD doesn't confirm, is treated as a genuine warning sign, particularly on a longer timeframe like a weekly chart.

Common Mistakes

• Treating MACD like a bounded oscillator and looking for "overbought" or "oversold" readings that don't actually exist on this indicator.
• Comparing MACD's raw number directly across two different-priced stocks, when the indicator isn't normalized like RSI is.
• Acting on every signal-line crossover equally, instead of weighing a fast signal-line cross very differently from a slower, heavier centerline cross.
• Assuming a real, working signal in one direction (like the Bitcoin bearish crosses above) means
  MYTH VS. REALITY

  MYTH "A MACD crossover is a clean buy or sell signal on its own."

  REALITY MACD is a lagging trend-confirmation tool that whipsaws frequently in sideways markets. This chapter's own Monte Carlo test showed only a +0.3 point edge over a random entry. The same real Bitcoin bearish crosses that called two real drops did not call the recoveries; treat it as confirmation alongside a bounded oscillator, not a standalone trigger.
  the same indicator is equally reliable in the opposite direction.

Watching This for Free

MACD is a free, default study on TradingView's free plan, Yahoo Finance, and effectively every retail broker's charting package, almost always shipped with the same 12, 26, 9 settings covered above.

1. 1Add MACD with the default 12, 26, 9 settings and watch the histogram, not just the two lines; it shows a shift starting before the actual crossover happens.
2. 2Treat a signal-line crossover as a faster, lighter signal, and a centerline crossover as a slower, heavier one. They aren't the same weight of evidence.
3. 3Never look for "overbought" or "oversold" on MACD: that's not what it measures, and there's no fixed line for it.
4. 4Pair it with a bounded oscillator like RSI or StochRSI for timing: MACD confirms the trend, the oscillator times the entry.
5. 5Check whether the same signal has actually been working in both directions recently, not just the direction you're hoping for.

Try It Yourself: Read a Real MACD Histogram

The fastest way to really understand MACD is to watch its histogram build in real time on a chart you already follow.

1. 1Add MACD (12, 26, 9) to any chart and find the last place the histogram crossed from negative to positive, or positive to negative.
2. 2Check what price did in the days right after that crossover: did it follow through, or fade like the Bitcoin bullish crosses above?
3. 3Look at whether MACD has crossed its own zero line recently, and compare how that felt different from a signal-line cross alone.
4. 4Pull up the same chart's RSI or StochRSI at that same moment and see whether it agreed with MACD or disagreed with it.
5. 5Note that disagreement, if you found one. Two indicators pointing different directions at once is exactly when caution matters most.

That's the real skill this chapter is teaching: MACD isn't telling you where price is on a scale, like RSI does; it's telling you whether the trend underneath price is speeding up, slowing down, or reversing, which is a genuinely different question with a genuinely different kind of answer.

Visualize It: What Moves the MACD Histogram

MACD is the gap between a fast and a slow EMA. The histogram is the gap between MACD and its signal line.

The Newsletter Publisher Who Built MACD by Hand

Gerald Appel ran a registered investment advisory and wrote a market newsletter, Systems and Forecasts, for decades. MACD grew out of that newsletter work in the 1970s — Appel wanted a way to track the relationship between two moving averages without redrawing charts by hand every day. He never patented it or tried to control how it spread; he just kept publishing about it, and traders kept copying it into their own charting software. By the time personal computers made plotting indicators trivial, MACD was already the default trend-momentum tool taught alongside RSI.

Cypher Guide: MACD Updated Bar by Bar, With the Histogram Flip Level

This guide fixes the vocabulary of MACD and then works the three nested calculations by hand: one full update step from the chapter’s numbers, the close that would flip the histogram, and the weights that show what the indicator really measures. All figures are hypothetical.

Mechanism one: one full update step. Start from the chapter’s state: EMA(12) = 182.40, EMA(26) = 178.65, signal = 2.90, so the MACD line is +3.75. A new close arrives. Each EMA moves by its own multiplier times the miss, the new gap is formed, and the signal takes 0.2 of the gap between the new MACD and the old signal.

Read across the row for a close of 182.40, equal to the old EMA(12). The fast average does not move, the slow one still rises, and the histogram shrinks from 0.85 to 0.458. A flat-to-down close can shrink the histogram even while MACD is positive. Each extra dollar of close adds 0.0798 to the MACD line and 0.0638 to the histogram, so the histogram is a damped response to price.

Mechanism two: what the histogram flip needs. The histogram is zero when the new MACD equals the old signal, because the signal moves only 0.2 of the way. Solving for the close gives the price at which momentum, in this sense, turns.

From a close of 182.40 the flip level is 175.23, a drop of 7.17 dollars, or 3.9 percent. The histogram can stay positive through a surprisingly large decline, which is one reason a histogram flip confirms a turn late.

The MACD weights add to zero, because each EMA's weights add to 100 percent. They are positive for recent closes and turn negative from 9 bars ago. MACD is therefore a measure of recent price against older price, not a price level, and a long flat price drives it to zero.

In plain words: MACD is a weighted comparison of recent closes with older closes, passed through a second smoother, so each line answers more slowly than the one before it. Education, not advice.

Closing Rundown: What MACD Measures and What It Cannot Tell You

Chapter 3 set out to show how MACD is built from two exponential averages and how its three parts are read. The one idea is that MACD measures how fast two averages are pulling apart or together, so it has no scale of its own and no overbought or oversold level.

• Three parts, each built on the last. The MACD line is the 12-day EMA minus the 26-day EMA, the signal line is a 9-period EMA of that line, and the histogram is the gap between the two. From EMAs of 182.40 and 178.65 with a signal of 2.90, the line is 3.75 and the histogram is 0.85.
• Each EMA gives the new close a weight of 2 ÷ (N + 1). A close of 190.00 lifted the MACD line from 3.75 to 4.078 and the histogram from 0.85 to 0.943, because the fast average moves more than the slow one.
• Two crossovers, two weights. A signal-line crossover is faster and more frequent, and matches the histogram flipping sign. A zero-line crossover is slower and means the fast EMA has crossed the slow one outright.
• No overbought or oversold. The raw value cannot be compared across markets. A MACD of 3.75 is 2.1 percent of a 178.65 slow EMA but 18.8 percent of a 20.00 slow EMA.
• The histogram turns late. From a close of 182.40 the histogram only reaches zero at 175.23, a drop of about 3.9 percent. A flip confirms a turn after much of it has happened.
• 12, 26 and 9 are conventions. They come from the old six-day trading week. Keeping them keeps your chart comparable with everyone else’s.

Chapter 4: Moving Averages — SMA, EMA, and the Golden Cross

RSI smooths gains and losses. MACD is nothing but two moving averages and the gap between them. Both are secretly built on top of the same tool, the one neither chapter has named outright. This chapter covers that tool directly: the moving average itself, and the single most-watched crossover in all of technical analysis.

Streamline View: How Moving Averages Are Built and Read

A moving average smooths a run of closing prices into one line that shows the direction of the trend. The rule that matters most is that every moving average looks backward, so it confirms a trend rather than calling the turn. The two ways it goes wrong are treating a crossover as a bottom call, and trading crosses mechanically in a sideways market where they reverse almost at once.

Second outcome. In a trending market, a golden cross that arrives with price at 110 still leaves a trend to ride. In a sideways market, the 50-day and 200-day can sit close together and cross, uncross and cross again within weeks, and each cross is a signal that does not last. The averages did not fail; the market simply offered no trend to confirm.

In plain words: a moving average shows where the trend has been, an EMA reacts faster than an SMA, and a golden or death cross confirms a move that is already well under way. Education, not advice.

History of Moving Averages: From Statistics to Stock Charts

The moving average began as a statistician’s tool for smoothing data, became an engineer’s tool for forecasting stock levels, and only then reached the charts of traders. The exponential version has the most traceable line from one field to the next.

• The EMA length and its percent are the same thing. The multiplier 2 ÷ (N + 1) gives 10% for N = 19, which is why Haurlan’s 10% Trend equals a 19-day EMA. Older sources talk in percents, newer ones in days, and the two convert by one line of algebra.
• The formula was built for cheap computing. It needs only yesterday’s value, so it was practical long before charting software. The same property is why it appears inside so many later indicators, such as MACD.
• The crossover rule is older than its famous names. A short average crossing a long one was published as a trading method in 1960, the same shape as the 50-day and 200-day cross taught in this chapter, only with shorter lengths. Its limits are old too: it works when a trend exists and whipsaws when it does not.

Who Built This, and Why

The moving average itself predates modern charting software by more than a century. R. H. Hooker was calculating "instantaneous averages" of economic data as early as 1901, and the term "moving average" was formally applied by G. U. Yule in 1909, describing Hooker's method in a statistics journal. The exponential version has a stranger origin story: Pete Haurlan, a rocket scientist at Caltech's Jet Propulsion Laboratory, applied exponential smoothing to stock prices in the early 1960s. He didn't call it an EMA. He called them "Trend Values," labeled by their smoothing percentage (what's now a 19-day EMA, he called a "10% Trend"). His aerospace background is exactly why the math works the way it does: an EMA only needs yesterday's average and today's price to update, not an entire stored history, a real constraint on 1960s computing that still shapes the formula today.

Two Ways to Average: SMA vs EMA

A simple moving average is exactly what it sounds like: add up the closing prices over N days and divide by N. Every day in the window counts equally, and old data simply falls off the back as new data comes in. An exponential moving average weights recent price more heavily, so it reacts faster and lags less: the tradeoff most other indicators in this module, including MACD, are built around.

Notice the EMA only moved from 48.20 to 48.53 even though the new close was a full $1.80 higher. It absorbed about 18% of that move and carried the rest of its own history forward. That's the entire mechanism in one example.

Common Periods, and What They're Used For

Shorter periods (5–20) track fast, near-term moves and are common on intraday charts. Medium periods, especially 50, are read as the trend traders are actually positioned around right now. The 200-day is the one everyone watches, on every asset class, not because 200 is mathematically special, but because so many other traders are watching the exact same line that it becomes a real, self-reinforcing level.

The Golden Cross and the Death Cross

These are the single most-referenced moving average signals in the entire industry. A golden cross is the 50-day moving average crossing above the 200-day, read as a bullish, longer-term trend change. A death cross is the reverse: the 50-day crossing below the 200-day, read as bearish.

A Recent Golden Cross Claim, Checked

On April 8, 2026, market commentary reported that the S&P 500 had formed a golden cross on its weekly chart (the 50-week moving average crossing above the 200-week), describing it as the first since the post-pandemic recovery. The claim does not hold up on the chart itself: a golden cross needs the 50-week line to have been below the 200-week line first, and through the 2022 and 2025 declines the 50-week line stayed above it. The useful habit is to check a reported crossover on the chart before trusting the headline.

Where This Signal Actually Fails

A backtest of the classic 50-day/200-day golden cross and death cross rule on the S&P 500 from 2003 through 2023 captured most of the major bull runs, including the recoveries after 2009 and 2020, but it also produced real whipsaw signals during sideways, range-bound stretches. The U.S. Global Jets ETF is a documented real example: during 2023–2024's choppy, directionless trading, it threw off several short-lived golden cross and death cross signals in a row, each one reversing almost immediately. The same signal that nails a real regime change can fire uselessly, over and over, in a market that isn't actually trending at all.

How Professionals Actually Use It

Moving averages are rarely a signal on their own — they're the trend filter everything else in this module gets read against. Price above a rising 50-day and 200-day is read as a genuine uptrend; RSI, StochRSI, and MACD signals that agree with that trend are trusted more, and signals that fight it are trusted less. Many traders also stack several moving averages at once (a "ribbon" of 10, 20, 50, and 200) and read how tightly bunched or spread out they are as its own read on trend strength.

Common Mistakes

• Treating a golden cross as a buy signal at the bottom, when price has typically already run 10–20% by the time the cross actually happens.
• Trading golden and death crosses mechanically in a sideways market, exactly the whipsaw pattern the Jets ETF example shows.
• Treating the 200-day as an exact price level instead of a zone: moving averages are watched for support and resistance areas, not precise lines.
• Using a slow SMA when the situation actually calls for a faster-reacting EMA, or vice versa, without knowing why the choice matters.

Watching This for Free

Every major free charting platform (TradingView, Yahoo Finance, and virtually every broker) plots SMA and EMA at any period for free, and most will flag a golden cross or death cross automatically the moment it happens.

1. 1Add both the 50-day and 200-day moving average to a chart you already follow.
2. 2Check whether the market is actually trending or chopping sideways first: crossovers are far less reliable in a range.
3. 3Treat a golden or death cross as confirmation of a trend already underway, not a call for the exact top or bottom.
4. 4Use the moving average as the trend filter for your other indicators; trust RSI, StochRSI, or MACD signals more when they agree with it.
5. 5Treat the 200-day as a zone, not a precise line, when judging support or resistance.

Try It Yourself: Spot the Last Real Crossover

The fastest way to really understand moving averages is to find a real one on a chart you already know.

1. 1Add the 50-day and 200-day SMA to any chart and scroll back to find the most recent golden or death cross.
2. 2Check how far price had already moved by the time the lines actually crossed.
3. 3Swap the 200-day SMA for a 200-day EMA on the same chart and compare how much earlier the EMA reacts.
4. 4Look for a stretch of sideways, range-bound price nearby and check whether any crossovers happened there, and whether they whipsawed.
5. 5Note what RSI, StochRSI, or MACD were doing at that same real crossover: did they agree with it or not?

That's the real skill this chapter is teaching: a moving average isn't a signal by itself so much as it's the lens everything else in this module gets viewed through — the trend that tells you whether a momentum reading is confirming the move or fighting it.

Visualize It: How Far Does One Close Move the 20 EMA?

An EMA is a running average that leans on recent prices. Set the length, yesterday’s EMA and today’s close to see how much weight the new close gets.

The Trader Who Made Trend-Following an Industry

Richard Donchian is usually called the father of trend following. In 1949 he launched Futures, Inc., one of the first publicly available commodity trading funds, built almost entirely on moving average crossovers and breakout rules, decades before "systematic trading" was a phrase anyone used. His simplest rule, buy when a short average crosses above a long one and sell on the reverse, is the direct ancestor of the Golden Cross taught in this chapter. Donchian ran his fund on hand-plotted charts; the logic hasn't changed since, only the plotting.

Cypher Guide: SMA and EMA Step Response, and How Late a Crossover Prints

This guide fixes the vocabulary of moving averages and then works two calculations by hand: how an SMA and an EMA of the same length answer a price jump, and how long a crossover takes to appear after a trend begins. All figures are hypothetical.

Mechanism one: the step response. Let price sit at 100 for a long time and then jump to 110 and stay there. A 10-bar SMA climbs in a straight line and arrives exactly 10 bars later. A 10-bar EMA climbs quickly at first and then creeps. The number to compare is the share of the 10-point jump closed after k bars.

Read across the row for k = 3. The EMA has closed 45.2 percent of the jump and the SMA 30.0 percent, so the EMA looks faster. Look at k = 10: the SMA has finished, while the EMA still has 13.4 percent to go. Both have the same average age of 4.5 bars. The EMA answers sooner and finishes later, which is a difference of shape, not of overall lag.

Mechanism two: how late a cross arrives. Let price fall by 0.25 per bar for 600 bars into a low of 100, then rise by 1 per bar. The table finds the first bar after the turn on which the short average moves above the long one.

Read across the row for 50 and 200. The SMA pair crosses on bar 45, when price is already 45 percent above its low of 100; the EMA pair crosses on bar 44, one bar sooner. Shorter pairs cross sooner. The slopes here are assumed (a steep rise of 1 percent of the low per bar), so the bar numbers show the order of magnitude, not a forecast, and are a separate example from the 10 to 20 percent rule of thumb used earlier in the chapter. Even so, every cross arrives after the turn is well under way, and in a jagged market the shorter pairs would also cross back and forth more often.

In plain words: an SMA and an EMA of the same length have the same average lag but different shapes, and a crossover of long averages is a delayed confirmation, not an early warning. Education, not advice.

Closing Rundown: What Moving Averages Confirm and Why They Lag

Chapter 4 set out to show how moving averages are built, how the two common kinds differ, and what the golden and death crosses can and cannot tell you. The one idea is that every moving average looks backward, so it confirms a trend that is already under way and does not call the turn.

• Two ways to average. An SMA adds the last N closes and divides by N. An EMA applies a multiplier of 2 ÷ (N + 1), about 18.2 percent for 10 periods and 9.5 percent for 20.
• Same lag, different shape. Both a 10-bar SMA and a 10-bar EMA have an average age of 4.5 bars. After 3 bars the EMA has closed 45.2 percent of a jump and the SMA 30.0 percent, but the SMA finishes in 10 bars while the EMA is still creeping.
• Shorter periods follow faster. Periods of 5 to 20 track near-term moves, 50 is the intermediate trend, and the 200-day matters because so many traders watch the same line.
• Crosses arrive late. A golden cross is the 50-day moving above the 200-day and a death cross is the reverse. Either appears after the turn is well under way, so it confirms and does not predict.
• Sideways markets break the rule. With no trend the 50-day and 200-day sit close together and can cross, uncross and cross again within weeks, and each signal reverses almost at once.
• Use it as context. The averages did not fail in a range; the market gave them no trend to confirm. Read crosses with price structure and other tools.

Chapter 5: Bollinger Bands — Trading the Width of Uncertainty

The last chapter built the moving average: the single line that smooths price into a trend. Bollinger Bands take that same line and wrap it in a second measurement most traders never think to ask for: how confident should you be that price stays near it? The answer changes constantly, and when the bands squeeze down to almost nothing, that answer is about to get tested.

Streamline View: How Bollinger Bands Are Built and Read

Bollinger Bands wrap a 20-day moving average in two lines set a multiple of the recent standard deviation above and below it. The rule that matters most is that the width shows how much price has been moving, so tight bands mean a quiet market and wide bands a restless one. The main way it goes wrong is reading a touch of the upper band as an automatic sell, when in a strong trend price can walk along the band for weeks.

Second outcome. If the same average were held but the standard deviation were half as large, at 1.23, the bands would run from 143.19 to 148.09, a width of 3.4%. The same 148.60 close would then sit above the upper band. A close outside tight bands after a squeeze is the kind of breakout the squeeze flags, though it does not say whether the move will hold.

In plain words: the bands show how much price has been moving around its average, a squeeze says a bigger move is due without naming a direction, and a touch of the band is a fact about spread, not an order to sell. Education, not advice.

History of Bollinger Bands: From Fixed Envelopes to Adaptive Bands

Bollinger Bands are the best-known descendant of a much older family, the trading bands drawn around a moving average. The family had existed for years as flat-percentage envelopes, alongside other band designs such as Keltner bands and Donchian bands, which draw their edges from ranges rather than from standard deviation. The change that made this member lasting was small: let the standard deviation set the width, and the bands begin to track the market’s own mood.

• The width is the idea. This is the main difference from the older designs. The older flat-percentage bands drew the same envelope in calm and wild markets. The newer bands move, so width itself became a signal: the squeeze exists only because the bands can pinch.
• The standard deviation is the population form. Bollinger divides by the number of closes, 20, and not by 19. On the chapter’s sample the figure is 2.45, where the other form gives 2.51, so platforms that choose differently will draw slightly different bands. The 20 and the 2 are the author’s defaults, and a chart that keeps them can be compared with the books and with other traders’ screens.
• The creator warns against the most common misuse. The chapter records his repeated warning that a touch of a band is not a signal by itself. It is a measure of how stretched price is against its own recent spread, and it needs a second tool to say what that stretch means.

Who Built This, and Why

John Bollinger developed Bollinger Bands in the early 1980s while working as a technical analyst, frustrated with the fixed-percentage "trading bands" popular at the time (bands drawn a flat 4% or 5% above and below a moving average, regardless of whether the market was calm or chaotic). Bollinger's insight was to let the bands breathe: instead of a fixed percentage, he measured how far price had actually been wandering from its own average, and used that wandering (statistical standard deviation) to set the width. A quiet market gets tight bands. A turbulent one gets wide bands. The indicator adapts to the market instead of asking the market to fit a fixed rule.

The Formula: A Moving Average With a Margin of Error

Three lines make up the indicator, all built from the same 20-day price series:

Standard deviation is just a measure of how spread out the last 20 closes are around their own average. Tight cluster of closes, small standard deviation, narrow bands. Closes scattered all over the place, large standard deviation, wide bands. Multiplying by 2 is Bollinger's own default, chosen because, in many samples of price data, roughly 88–89% of closes fall inside a 2-standard-deviation band, leaving a meaningful minority of closes to push outside it and flag something unusual happening.

The Squeeze: When the Bands Tell You to Pay Attention

Because the bands are built from standard deviation, they don't just show where price is; they show how much price has been moving at all. When volatility dries up, the 20-day standard deviation shrinks, and the upper and lower bands pull in toward the middle line. Bollinger called this pattern the squeeze, and it's arguably the single most useful signal the indicator produces: a tight squeeze means the market has gone unusually quiet, and low volatility has a strong tendency to resolve into high volatility. The squeeze doesn't tell you which direction the breakout will go. Only that one is statistically overdue.

Riding the Band: Why Touching the Upper Band Isn't Automatically "Overbought"

New traders often treat a touch of the upper band the way they'd treat RSI crossing 70, as a sell signal. That's a mistake Bollinger himself has repeatedly warned against. In a strong, healthy uptrend, price can walk right along the upper band for weeks, closing above or against it session after session, precisely because the trend is strong enough to keep pushing standard deviation outward faster than price can catch up. Reading a band touch as automatically overbought means

missing the strongest trends the indicator will ever show you.

A Real, Very Recent Example: The Creator Trades His Own Signal

On May 7, 2026, a crypto news article reported something unusual: John Bollinger himself said one of his investment fund's proprietary trading models "had flipped positive on bitcoin and taken a position accordingly." This wasn't a media pundit relaying the indicator; it was the person who invented it, acting on it, in public, in real time. Bitcoin had closed above its upper daily Bollinger Band the previous Wednesday, May 6, 2026 (only the second such close since mid-January) and that breakout followed the tightest Bollinger Band reading Bitcoin had ever recorded, a textbook squeeze.

By the time the news article published on May 7, Bitcoin was trading at $80,484, with the upper Bollinger Band itself reading $81,549, price and band nearly touching. Bitcoin was up roughly 9% over the trailing 30 days, but still down 36% from its October 2025 peak of $126,000. The article also noted a separate on-chain indicator, MVRV, reaching levels comparable to those seen before Bitcoin's late-2024 surge toward $100,000.

The Big Limitation: Bands Don't Predict Direction, and They Don't Work Alone

Bollinger Bands measure dispersion, not momentum or trend direction. A price sitting on the upper band could be a powerful breakout or the exhausted final gasp of a rally — the bands alone can't tell you which. That's precisely why standard investor education guidance states plainly that the indicator is "not intended to be used on its own," and recommends pairing it with a momentum oscillator, exactly the kind of tool built in the earlier chapters of this module. If price pushes to a new high near the upper band while RSI or MACD fails to confirm with a new high of its own, that divergence (covered in Chapter 1 and Chapter 3) is one of the most reliable warnings the two tool families can give together.

How Professionals Actually Use It

Desks rarely trade a single band touch in isolation. The squeeze is used as a volatility alarm: a reason to start watching a name closely, not a trade signal by itself. Band walks (price riding the upper or lower band through a strong trend) are read as trend-strength confirmation, not exhaustion, unless a momentum oscillator disagrees. And bandwidth itself (how wide the bands are as a percentage of price) is tracked as its own mini-indicator: historically narrow bandwidth for a given stock or asset is the closest thing this indicator has to a genuine "something is about to happen" flag.

Common Mistakes

• Treating an upper-band touch as an automatic sell signal, and getting run over by a strong trend that keeps walking the band for weeks.
• Trading every squeeze the moment it appears, instead of waiting for the actual breakout: direction and confirmation, not just tightness, is the entry signal.
• Using Bollinger Bands as a standalone system, ignoring the common guidance to pair it with a momentum tool from earlier in this module.
• Forgetting the bands are recalculated fresh every day from the trailing 20-day window: a level that looked like resistance last week may not still be the band today.

Watching This for Free

Bollinger Bands are a default overlay on TradingView's free tier, Yahoo Finance charts, and essentially every retail broker's charting package. No paid data feed required. John Bollinger's own site also publishes free educational material directly from the indicator's creator.

• 1Add a 20-period, 2-standard-deviation Bollinger Band overlay to any free charting platform; it's usually a one-click default.
• 2Watch for bandwidth compressing to multi-month lows: that's your squeeze alert, not a trade signal yet.
• 3Wait for a confirmed close outside a band, in either direction, before treating the squeeze as resolved.
• 4Cross-check the breakout against RSI, StochRSI, or MACD from earlier chapters: agreement raises confidence, disagreement is a warning.
• 5If price is walking a band through a strong trend, don't treat the touch itself as a reason to exit. Watch for the momentum tool to roll over first.

Try It Yourself: Spot a Real Squeeze

1. 1Pull up any liquid stock or ETF on a free charting platform and add a 20-period, 2-standard-deviation Bollinger Band overlay.
2. 2Scroll back through the last 6–12 months and find the single tightest stretch where the bands pinch closest together.
3. 3Note what happened in the 1–3 weeks after that squeeze: which direction did price break, and how far did it travel?
4. 4Check whether RSI or MACD was already leaning in that breakout direction before the bands confirmed it.
5. 5Find one recent band walk (a stretch where price rode the upper or lower band for multiple sessions) and note whether treating it as "overbought" or "oversold" would have cost you the rest of the move.

Bollinger Bands close the loop this module opened with RSI: momentum tools tell you how hard price is moving, and Bollinger Bands tell you how unusual that movement is relative to its own recent history. Neither one, alone, is a complete system, which is exactly the point their own creator keeps making, including with real money, in public.

Visualize It: Build Bollinger Bands

The bands are an average plus and minus a number of standard deviations. Set the inputs and see the width and where price sits.

The Analyst Who Wanted Bands That Adapted to Volatility

John Bollinger was a technical analyst and options trader who kept running into the same problem with the fixed-width price envelopes traders used in the 1970s: a band that worked in a calm market was useless in a volatile one. In the early 1980s he started scaling the bands to a security's own recent standard deviation instead of a flat percentage, so they widened and narrowed with the market automatically. He trademarked the name but taught the method freely, and it's now one of the handful of indicators recognizable by name to people who've never placed a trade.

Cypher Guide: Bollinger Bands From Raw Closes, With %b and Bandwidth

This guide fixes the vocabulary of Bollinger Bands and then works the calculation from the raw closes: the mean, the sum of squared deviations, the population standard deviation, the bands, and the two derived readings %b and bandwidth. A rolling step and a multiplier comparison follow. All figures are hypothetical.

Mechanism one: from raw closes to bands. The chapter lists twenty closes. The steps are: add them, divide by 20, square each deviation, add the squares, divide by 20, take the square root. Doing it this way shows where the chapter’s rounded figures come from.

Read across the %b line. A close of 148.60 sits at 0.80 of the way from the lower band to the upper band: inside the envelope, in its top fifth, but not at the edge. The chapter’s figures all agree with the raw calculation. Bandwidth of 6.7 percent is the number to compare with later readings to judge a squeeze.

The new close of 150.80 is above the old upper band of 150.54, yet its %b is only 0.97 against the new bands, because the bands moved up as well. The oldest close left the window at the same time, and a low value leaving a window raises the mean. A touch is always measured against bands that the same bar has changed.

Mechanism two: the multiplier. Changing k stretches the envelope in proportion. The last column is the share of data any distribution must keep inside k standard deviations (Chebyshev’s bound). The chapter’s 88 to 89 percent is an empirical figure for price series; a normal curve would give 95.4 percent for k = 2.

Read across the k = 3 row. The envelope runs from 138.29 to 153.00, so a close of 148.60 is nowhere near the edge. A multiplier is a choice about how rare a touch should be, so the same price can be a band touch at k = 1.5 and ordinary at k = 3.

In plain words: Bollinger Bands are a moving average with a margin that grows when the last 20 closes are scattered, so %b and bandwidth are simply position and width measured against that margin. Education, not advice.

Closing Rundown: What Bollinger Bands Measure and How to Read Them

Chapter 5 set out to show how Bollinger Bands are built from a moving average and a measure of spread, and how to read their width and position. The one idea is that the bands measure how much price has been moving, not which way it will go.

• One average, one spread. The middle band is the 20-day SMA and the outer bands sit 2 standard deviations away. For the chapter’s 20 closes that gave 145.64, 150.54 and 140.74, with a width of 6.7 percent.
• %b is position, bandwidth is width. The last close of 148.60 sat at (148.60 − 140.74) ÷ (150.54 − 140.74) = 0.80, inside the bands and in their top fifth.
• A squeeze flags a coming move, not a direction. Quiet markets shrink the standard deviation. Halving it to 1.23 would narrow the width from 6.7 to 3.4 percent and put the same close above the upper band.
• A touch is not a sell. In a strong trend price can ride the upper band for weeks. A band walk reads as strength unless a momentum tool such as RSI or MACD disagrees.
• A touch is measured against moving bands. Adding a close of 150.80 and dropping the oldest value gave a %b of only 0.97, because the bands rose too.
• The multiplier is a choice. Width grows in direct proportion to k. At k = 3 the envelope ran from 138.29 to 153.00, so the same close was ordinary there.

Chapter 6: On-Balance Volume (OBV) — Volume Leads Price

Price has been the only raw material this module has used so far. OBV throws that out and builds from something else on the chart that most beginners ignore completely: how many shares actually changed hands. The idea behind it is simple and a little unsettling: that big money tends to start buying or selling before the price fully shows it, and volume is where that shows up first.

Streamline View: How On-Balance Volume Is Built and Read

On-Balance Volume (OBV) is a running total of volume that adds the day’s volume when the close rises and subtracts it when the close falls. The rule that matters most is that only the direction of the line and its agreement with price carry information; the number itself means nothing. The main ways it goes wrong are reading the absolute level, and trusting one volume spike as if it were a trend.

Second outcome. Suppose price makes a lower low at 49.90 after an earlier low of 50.20, while OBV at the second low is 900,000 against 300,000 at the first. That is a bullish divergence: sellers moved price lower with less volume behind them. It is a clue to watch, not a buy order. For scale, the GameStop session on January 25, 2021 saw more than 175 million shares against a 30-day average of 29.8 million, a ratio of 5.9 times, yet OBV would not have shown how large the move would be.

In plain words: OBV keeps a running tally of volume by the direction of the close, its slope and its divergence from price are what count, and its level is not comparable between stocks. Education, not advice.

History of On-Balance Volume: From Continuous Volume to 1963

OBV is a bookkeeping idea older than its famous name: sum volume with a sign set by the direction of the close. Its history follows one author, Joseph Granville, from an earlier technique he renamed, through a newsletter that gave it an audience, to a public record of market calls that has little to do with the maths.

• The claim is a hypothesis, not a law. “Volume precedes price” is the 1963 premise. The line can show pressure building, but nothing in its history shows that it measures the size of a coming move.
• The running total has no natural zero. Because it was built as a ledger, the level depends on where counting began, and the fact that it is cumulative applies to every platform: two charts that start on different dates show different numbers for the same days. That is why a slope or a divergence is read and the raw number is ignored.
• The tool and the author are separate questions. Granville’s calls drew heavy attention in 1980 and 1981 and his later forecasts were uneven, yet OBV is still on every chart. Judge the line on its own record against price, not on the standing of its inventor, and weigh the same caution for any indicator tied to a famous name.

Who Built This, and Why

Joe Granville introduced On-Balance Volume in his 1963 book Granville's New Key to Stock Market Profits. His central claim, distilled into a single phrase that's still quoted today, was that "volume precedes price" — that a rising tide of buying or selling pressure shows up in trading volume before it fully shows up in the price itself, giving attentive traders an early read on where a stock is headed.

The Formula: A Running Total That Only Cares About Direction

OBV doesn't use volume's actual size beyond a simple add-or-subtract rule. It's a cumulative running total, adjusted every day based on whether the close went up, down, or stayed flat:

That's the entire calculation. There's no lookback window, no smoothing, no averaging: just a running ledger of volume added on up days and subtracted on down days, all the way back to whenever you started counting.

Reading OBV: It's the Slope and the Divergence That Matter, Not the Number

Charting-platform technical documentation is explicit that OBV's absolute value means nothing on its own; the number itself isn't even shown on many charting platforms. What matters is the shape of the line: a rising OBV suggests buying pressure that can precede higher prices, a falling OBV suggests selling pressure that can precede lower prices, and OBV moving in the opposite direction of price (divergence) is the single most-watched signal the indicator produces. Bullish divergence is OBV making a higher low while price makes a lower low; bearish divergence is OBV making a lower high while price makes a higher high.

A Real Example: The Loudest Volume Signal in Modern Market History

GameStop's January 2021 squeeze is the clearest real-world illustration of Granville's "volume precedes price" idea that modern markets have produced. The stock started January 2021 trading around $17.25. On January 13, it surged more than 50% in a single session. By January 22, call-option buying had exploded and market makers were forced into large-scale delta-hedging, triggering the first major gap up. On January 25 alone, more than 175 million shares changed hands, against a trailing 30-day average of just 29.8 million, nearly six times normal turnover. By the January 28 intraday peak of $483, the stock had gained over 2,700% in under a month.

How Professionals Actually Use It

Desks rarely trade a single OBV divergence in isolation; it's used as a confirmation layer, checking whether a price breakout is backed by real volume conviction or is happening on thin, unconvincing turnover. A price making new highs on a rising OBV is read as a healthy, well-supported trend; a price making new highs while OBV stalls or falls is read as a warning that the move may be running on fumes. Institutional desks also watch abnormal single-day volume spikes, exactly like GameStop's 175-million-share day, as a signal that something structurally unusual (short covering, options hedging, a sudden shift in sentiment) is happening beneath the surface.

Common Mistakes

• Reading the absolute OBV number as meaningful on its own, when only its slope and divergence from price actually carry information.
• Treating a single volume spike as a reliable signal, when standard charting documentation itself warns that spikes can throw off the indicator with a sharp, one-off jump.
• Using OBV as a standalone trading system instead of a confirmation tool alongside price action and the momentum indicators covered earlier in this module.
• Ignoring that OBV is calculated purely from closing prices: a volatile intraday session that closes flat contributes nothing to the running total, even if a lot happened during the day.

Watching This for Free

OBV is a standard, no-cost overlay on TradingView's free tier, Yahoo Finance charts, and most retail broker platforms, typically found in the same indicator list as RSI and MACD from earlier in this module.

• 1Add OBV as a sub-chart beneath price on any free charting platform; it's usually a one-click addition from the indicator list.
• 2Ignore the actual number on the OBV axis entirely. Focus only on whether the line is rising, falling, or flat.
• 3Compare OBV's highs and lows against price's highs and lows, watching specifically for divergence between the two.
• 4Flag any single-day volume figure that's several multiples of the recent average: that's the kind of signal that showed up before GameStop's real 2021 move.
• 5Use OBV to confirm a breakout from Chapter 5's Bollinger Bands or a moving-average cross from Chapter 4, rather than trading OBV in isolation.

Try It Yourself: Spot a Real Divergence

1. 1Pull up any liquid stock on a free charting platform and add an OBV sub-chart beneath the price panel.
2. 2Scroll back through the last 6–12 months and find a stretch where price made a new high or low.
3. 3Check whether OBV made a matching new high or low at the same time, or whether it diverged.
4. 4Note the single highest-volume day in that window and what price did on that specific day and the days right after it.
5. 5Decide, honestly, whether the OBV signal would have told you anything useful before the price move happened, or only after — that's the real test of the tool.

Every indicator up to this chapter has been built entirely from price. OBV is a reminder that price is only half of what a chart is actually recording, and that the other half, volume, sometimes moves first.

Visualize It: One Day of On-Balance Volume

OBV adds the day’s volume on an up close and subtracts it on a down close. Set the inputs and see the new value.

The Newsletter Writer Who Said Volume Leads Price

Joseph Granville was one of the most famous, and most controversial, market forecasters of the 1960s and 70s. His 1963 book "New Key to Stock Market Profits" introduced On-Balance Volume on the premise that volume moves before price does. Granville's market calls later became erratic and his reputation took a hit, but OBV itself survived his own credibility problems intact because the underlying logic held up independently of who was making the forecasts. It's a reminder that an indicator's usefulness and its inventor's track record are two separate questions.

Cypher Guide: On-Balance Volume as a Ledger, and a Size-Weighted Variant

This guide fixes the vocabulary of On-Balance Volume and then rebuilds the running total day by day, showing what the rule keeps and what it throws away. A comparison with a variant that also uses the size of the price move follows. All figures are hypothetical, and the ten days are chosen so that they reproduce the chapter’s walkthrough values.

Mechanism one: the ledger. The table gives ten closes and volumes. The OBV column uses the rule exactly: add the day’s volume on an up close, subtract it on a down close, change nothing on a flat close. The values on days 2, 3, 4, 7 and 10 match the chapter. The last column adds a variant described below.

Read across the row for day 7. Price rose 0.80 and volume was 2.0M, and OBV added the full 2.0M. Had the close risen only 0.01, it would still have added 2.0M. The rule sees up or down and never how far. Volume traded on days 2 to 10 totals 9.6M, and the net of 5.9M is 61 percent of it.

Mechanism two: a variant with size. Price-volume trend adds volume × (close ÷ previous close − 1), so a 0.01 move adds almost nothing and a 0.80 move adds a lot. On the same ten days the variant ends at 0.073M against 5.9M for OBV; the two use different units and are compared by shape, not by level.

Read across both lines. A heavy-volume day with a small move is a large step for OBV and a tiny step for the variant. Neither is more correct; they answer different questions. OBV asks who carried volume. The variant asks who carried volume per unit of price progress.

In plain words: OBV is a running ledger of volume signed by direction, so only its slope and span carry information, and it ignores how large each price move was. Education, not advice.

Closing Rundown: What On-Balance Volume Tracks and What It Ignores

Chapter 6 set out to show how On-Balance Volume turns daily volume into a running ledger and how to read it. The one idea is that only the direction of the line and its agreement with price carry information, while the number itself means nothing.

• Three rules, one ledger. Add the day’s volume on an up close, subtract it on a down close, and change nothing on a flat close. In the Streamline View example, day 4 had 2,000,000 shares traded and added nothing because the close did not change.
• The level is arbitrary. OBV starts wherever the count began and cannot be compared between stocks. Read the slope and the span of the line, not the figure.
• Divergence is the main clue. A lower low in price at 49.90 against 50.20, with OBV at 900,000 against 300,000, is a bullish divergence. It is something to watch, not a buy order.
• Size of move is ignored. A close up 0.80 and a close up 0.01 both add the full volume. A variant that scales volume by the percentage move adds only 1.58 percent of it on day 7.
• Volume can lead price, not magnitude. A session of more than 175 million shares against a 30-day average of 29.8 million was 5.9 times normal. OBV can show the pressure building but cannot size the move.
• Net versus total. Over days 2 to 10, 9.6M shares traded but the net was 5.9M, or 61 percent of the total.

Chapter 7: Average True Range (ATR) — Measuring Volatility Itself

Chapter after chapter, one question has driven the indicators in this module: which way is price going? ATR doesn't answer it at all. It doesn't care about direction. It answers a completely different question: how much is this thing actually moving? That turns out to be one of the most practically useful numbers in trading, because it's the basis for how many professionals decide where to place a stop-loss and how large a position to take in the first place.

Streamline View: How ATR Is Built and Read

Average True Range (ATR) measures how much a market typically moves in a day, gaps included, and says nothing about direction. The rule that matters most is that it is used to size a stop and a position once a direction has been chosen elsewhere. The usual ways it goes wrong are freezing a stop distance set in a calm period, and assuming a stop guarantees an exit price when a gap can skip right past it.

Second outcome. Suppose news arrives overnight and the stock opens at $90.00 instead. The $93.94 stop is a market order once it triggers, so it fills near the open, a loss of about $10 per share against the planned $6.06. On a 33-share position that is about $330 against a $200 budget. ATR cannot prevent that. Only a smaller position can.

In plain words: ATR is the typical daily size of a market including gaps, it sets how far to put a stop and how many shares to hold, and it never tells you which way to trade. Education, not advice.

History of ATR: A Gap-Proof Range and What Followed

ATR answers one question: how far does this market usually travel in a day? The question is older than the indicator. Wilder’s contribution was a precise answer that survives gaps, and the later history of the tool is a history of other people borrowing that answer for stops, channels and position size.

• It was built for markets that gap. Wilder had commodities in mind, where limit moves and overnight jumps are routine. A range that skips gaps would have been wrong for those markets, and it is still wrong for any stock that opens far from its last close.
• The smoothing is Wilder’s, not the common EMA. His weight of 1 ÷ 14 is slower than the usual 2 ÷ (14 + 1) used for a standard EMA. Two platforms that use different smoothing show different ATR values for the same bars.
• The offspring explain its limits. Keltner Channels, SuperTrend and the Chandelier Exit all hang a distance from price by a multiple of ATR. They inherit its blind spot: they size normal days, and a gap past the stop is not something any of them can prevent.

Who Built This, and Why

J. Welles Wilder, the same technical analyst who developed RSI back in Chapter 1, introduced Average True Range in the same 1978 book, New Concepts in Technical Trading Systems. He designed it specifically with commodities in mind, markets that could gap sharply overnight or hit daily limit moves that a plain high-minus-low range would completely miss.

The Formula: Why "True" Range Isn't Just High Minus Low

Wilder's insight was that a stock's real range on a given day isn't just today's high minus today's low: if the stock gapped up or down overnight, that gap is real movement too, and a simple high-low range ignores it completely. True Range is defined as the largest of three separate measurements:

ATR itself is simply a 14-period smoothed average of that True Range value, using the same style of smoothing Wilder used for RSI: New ATR = [(Prior ATR × 13) + Today's True Range] ÷ 14. The first ATR value is just a plain average of the first 14 True Range readings; every one after that uses the smoothing formula.

Using ATR for Stops and Position Size, Not Direction

Because ATR is non-directional, it's almost never used to decide whether to buy or sell; it's used to decide how to size and protect a trade once a direction is already chosen from another tool. The standard approach places a stop a multiple of ATR away from the entry price:

Position size then adjusts to match: if a trader's risk budget is $200 per trade and the stop sits $6.06 away from entry, that caps the trade at roughly 33 shares before costs: a wider ATR-based stop on a more volatile stock automatically means a smaller position, keeping the dollar risk consistent across very different tickers.

A Real, Dated Example: When Volatility Itself Made Headlines

The clearest real-world illustration of what ATR is measuring doesn't come from a single stock; it comes from the broad market's own volatility gauge. In April 2025, a wave of U.S. tariff announcements and the retaliatory measures that followed triggered a sharp global selloff. The CBOE Volatility Index (VIX), a close cousin of ATR that measures expected volatility through options prices rather than price ranges, closed at 52.33 on April 8, 2025, the highest level markets had seen since the COVID-19 crash of March 2020, when the VIX closed as high as 82.69. Between mid-February and early April 2025, the S&P 500 Growth Index fell more than 22%.

How Professionals Actually Use It

Beyond stops and position sizing, desks watch ATR expanding or contracting as a read on the overall character of a market. A breakout that occurs alongside expanding ATR is read as more credible (real participation is driving real range expansion) while the same breakout on flat or shrinking ATR is treated with more suspicion. Some trend-following systems, like the well-known Chandelier Exit, build the entire trailing-stop mechanism directly on top of ATR, defaulting to a 22-period lookback and a 3× multiplier.

Hall of Shame for Stops: The Biggest Gaps

ATR sizes a stop from the normal daily range. The cases below are the days when the stock did not trade through the range at all. News arrived while the market was closed, and the stock opened far below where any stop could have been.

Common Mistakes

• Freezing an ATR-based stop distance and never recalculating it, so a stop set during a calm period becomes far too tight once volatility genuinely expands.
• Treating a stop as a guaranteed exit price: gaps, thin liquidity, and trading halts can all make the actual fill different from the planned distance.
• Comparing raw ATR values across two differently-priced stocks directly, when ATR is an absolute dollar figure, not a percentage: a $5 ATR means something very different on a $20 stock than on a $500 stock.
• Using ATR to decide direction at all, when it's purpose-built to measure magnitude of movement, not which way that movement is likely to go.

Watching This for Free

ATR is a standard, free indicator on TradingView, Yahoo Finance, and virtually every retail broker's charting platform, usually listed right alongside RSI and MACD from earlier in this module.

• 1Add ATR as a sub-chart beneath price on any free charting platform; the 14-period default is standard and rarely needs changing.
• 2Note the current ATR value in dollars, then convert it to a rough percentage of the current price for an apples-to-apples read on volatility.
• 3Use a 1.5–3× ATR multiple to set a stop distance on any trade idea generated by another indicator in this module.
• 4Size the position so the dollar distance to that stop matches a fixed risk budget, rather than trading a flat share count on every ticker.
• 5Recalculate ATR periodically rather than freezing it indefinitely, especially around news events, earnings, or macro shocks like April 2025's tariff selloff.

Try It Yourself: Size a Stop With Real ATR

1. 1Pull up any liquid stock on a free charting platform and add the 14-period ATR indicator beneath the price panel.
2. 2Note the current ATR value and multiply it by 2 to get an illustrative stop distance.
3. 3Pick a hypothetical $500 risk budget and calculate how many shares that stop distance would allow you to buy.
4. 4Scroll back to a known volatile period for that stock (an earnings date, a market-wide event) and see how much ATR expanded compared to normal.
5. 5Recalculate the same stop and position size using that expanded ATR, and compare how differently the two trades would have been sized.

ATR doesn't tell you where price is going — every other indicator in this module has tried to answer that question in its own way. What ATR tells you is how much room to give a trade once you've already decided, which turns out to be just as important as being right about direction in the first place.

Visualize It: Build One ATR Step

True range is the biggest of the day range and the two gaps from the prior close. Wilder smoothing then blends it into the average.

The Same Book, a Different Problem to Solve

ATR comes from the same 1978 Wilder book as RSI (Chapter 1), but it was solving a completely different problem: how do you measure volatility on a security that gapped overnight, so a simple high-minus-low range understates how far it actually moved? Wilder's “true range” fix (taking the largest of three possible ranges, including gaps) became the basis for ATR and, later, for dozens of other volatility-based tools built by other analysts, including Keltner Channels and SuperTrend elsewhere in this module.

Cypher Guide: True Range to ATR, the Gap Day Fade, and Stop Sizing

This guide fixes the vocabulary of ATR and then works the calculation from raw highs, lows and closes: fourteen true ranges, the seed average, a gap day entering the recursion, and the way that day fades out. A look at stop distance and position size closes the guide. All figures are hypothetical.

Mechanism one: seed and recursion. The table builds fourteen true ranges from hypothetical bars, starting from a previous close of 100.0. Each row finds the three candidates and keeps the largest. The seed ATR is their plain average. Then a gap day is fed through the recursion.

Read across bars 4 and 6. On bar 4 the plain range of 2.8 is the largest of the three, so TR equals it. On bar 6 price sits well above the previous close of 100.6, so column B (2.8) beats the plain range (2.5) and TR takes B. The gap day lifts ATR from 2.39 to 2.60, a rise of 8.7 percent from a TR that was 2.2 times the average.

The excess shrinks by 13 ÷ 14 each bar, so half of it is gone after about 9 bars and most of it after 40. One large bar therefore affects stops for weeks, even though the market has been calm since. Different smoothing changes this: on the chapter’s numbers (old ATR 2.80, new TR 6.00) Wilder gives 3.029, while an EMA with a = 2 ÷ 15 gives 3.227 and reacts faster.

Mechanism two: stop and size. A stop at k × ATR below entry fixes the risk per share, and the position size is the budget divided by it. The table uses the chapter’s ATR of 3.03 and a higher ATR of 5.00 after a burst of volatility, a risk budget of 200 dollars and an entry of 100.

Read across the row for k = 2. The chapter’s figure of 6.06 dollars per share allows 33 shares, matching the chapter. If ATR rises to 5.00, the same rule allows only 20 shares. Size shrinks as volatility grows, which keeps the dollar risk near 200 but does not remove gap risk: a close-to-open gap of 12 dollars costs 12 per share, not 6.06.

In plain words: ATR is a slow average of the largest of three daily distances, so a gap enters at full size and leaves gradually, and it sizes risk rather than predicting direction. Education, not advice.

Closing Rundown: What ATR Measures and How It Sizes Risk

Chapter 7 set out to show how Average True Range measures the typical daily size of a market, gaps included, and how that size is put to work. The one idea is that ATR says how far, never which way, so it sizes a stop and a position once a direction has been chosen elsewhere.

• True range takes the largest of three. It is the greatest of high − low, |high − previous close| and |low − previous close|. In the gap example the plain range was $2.00 but the true range was $6.00.
• Wilder smoothing keeps a gap for weeks. The update is (previous ATR × 13 + new TR) ÷ 14. A TR of 6.00 lifted ATR from 2.80 to 3.03, and the excess fades by 13 ÷ 14 a bar, so half remains after roughly 9 bars.
• Stop distance is a multiple of ATR. A long stop is entry − k × ATR, with k commonly 1.5 to 3. At 2 × 3.03 from a $100.00 entry the stop sits at $93.94.
• Size shrinks as volatility grows. Shares = budget ÷ risk per share. A $200 budget allows 33 shares at $6.06 of risk, but only 20 when ATR rises to 5.00.
• A stop is not a guaranteed price. It becomes a market order once triggered. An open at $90.00 against a $93.94 stop meant a loss near $10 per share, about $330 on 33 shares.
• Smoothing choice changes the reaction. On the same inputs Wilder gives 3.029 while an EMA with a = 2 ÷ 15 gives 3.227, so it reacts faster.

Chapter 8: VWAP — The Price Institutions Actually Get Graded On

Telling you what to do next has been the job of the indicators in this module up to this point. VWAP has a different job: it grades what a trader already did. It answers a narrow, unglamorous question: "was my average execution price actually good, compared to how the whole market traded today?" And that narrow question turned out to be so useful that VWAP became one of the most-watched lines on every institutional trading desk's screen, and eventually a popular retail tool too.

Streamline View: How VWAP Is Built and Read

Volume-Weighted Average Price (VWAP) is the average price of a session in which each price counts in proportion to the volume traded there. The rule that matters most is that it is a benchmark for grading an execution, not a forecast. The usual ways it goes wrong are using the session line on a chart that spans many days, where it resets every morning, and treating every touch of it as a bounce.

Second outcome. If the third bar had traded only 2,000 shares instead of 12,000, the same prices would give a VWAP of 100.60, pulled toward the heavier early bars. The line follows where the volume was, not just where the price went. This is why the benchmark must be computed from the day’s own volume.

In plain words: VWAP is the day’s volume-weighted average price, institutions use it to grade fills, and a chart line that starts again each session is a reference for the day and not a prediction. Education, not advice.

History of VWAP: From a 1984 Order to Anchored Lines

VWAP began as a way to answer a client’s question after a large trade: was the price fair for the day? It was a measuring stick before it was a chart line. Its history runs from one fund’s order, through an academic paper, into broker products, and only later onto the charts of individual traders.

• It is a report card first. Born as a test of whether an order was filled fairly, VWAP measures the past. Its use as a live support line came later and is a borrowed purpose, which is why it is not a forecast.
• Two broker products grew from it. A guaranteed VWAP order has the broker take the slippage risk for a higher commission. An algorithmic VWAP order slices the order in proportion to expected volume. Both aim to land on the benchmark, not to beat it.
• The anchor is a later idea. Session VWAP resets by design. Anchored VWAP, developed in the mid-1990s, was added so that one line could follow a catalyst for weeks, and Brian Shannon later popularized it among traders.

Who Built This, and Why

The first documented VWAP execution is credited to James Elkins, a trader at the New York agency brokerage Abel Noser, who used it in 1984 to execute an order for the Ford Motor Company pension fund. The idea gained real academic weight in 1988, when three researchers formalized it in a peer-reviewed finance paper studying the true transaction costs of trading on the New York Stock Exchange. From there it became the industry-standard benchmark for judging whether a large institutional order was executed well or poorly.

The Formula: A Running Average, Weighted by How Much Actually Traded

VWAP is a cumulative, volume-weighted average price, reset at the start of every trading session. Most charting platforms compute it bar by bar using the "typical price", the average of that bar's high, low, and close:

Every bar's typical price gets multiplied by that bar's volume, added to a running total, and divided by the running total of volume itself, so a high-volume bar pulls VWAP toward its price much harder than a quiet, low-volume bar does.

Reading VWAP: Institutional Benchmark, Retail Support/Resistance

Institutions use VWAP almost exactly the way it was designed: as a report card. A large pension fund or mutual fund trying to buy a big position wants its average fill price to land at or below the day's VWAP; a fund selling wants its average fill at or above it. Some brokers even guarantee VWAP execution as a service, with the gap between a trader's actual fill and VWAP tracked as "VWAP slippage", a real, measured performance number that execution desks are judged on.

Retail traders borrowed the same line for a different purpose: reading it as dynamic intraday support and resistance. Price trading above VWAP is read as a sign of net bullish control for the session; price below is read as net bearish control. A common setup watches for price to break above VWAP, pull back down to retest it, and hold, treating that retest as confirmation the move has real backing.

Anchored VWAP: Starting the Clock Wherever You Want

Standard VWAP resets automatically at the start of every session, which makes it useless for anything beyond a single day. Anchored VWAP (AVWAP) fixes that by letting a trader pick their own starting point (an earnings announcement, a big gap, a major news event, a swing low) and calculates the same running volume-weighted average from that point forward instead of from the day's open. A trader watching how price behaves relative to the VWAP anchored at a stock's post-earnings gap, weeks later, is really asking: "is the market, on average, still validating that move?"

A Real, Dated Example: The Origin Story Is the Track Record

Unlike most indicators in this module, VWAP's most important real-world proof isn't a single dramatic chart; it's the fact that the entire modern institutional trading industry adopted it as a formal standard. James Elkins' 1984 Ford Motor Company pension fund execution at Abel Noser wasn't a one-off curiosity; the 1988 academic paper that followed it became a foundational reference for measuring transaction costs, and VWAP has been the default execution benchmark on trading desks for decades since. Brokers built entire algorithmic execution products around matching it, and asset managers use VWAP slippage as a routine, contractual performance metric on real institutional order flow, every trading day.

How Professionals Actually Use It

Execution desks run VWAP algorithms that slice a large order into smaller pieces spread across the session, timed to match expected volume patterns, specifically to keep the average fill price close to the benchmark and avoid moving the market against themselves. Day traders use VWAP as a session-long bias filter (favoring long setups only while price holds above it, short setups only while it holds below) and use anchored VWAP from a specific catalyst to judge whether that catalyst's move is still intact weeks later.

Common Mistakes

• Using standard session VWAP on a multi-day or swing-trade chart, where it resets every day and stops being a meaningful reference line.
• Treating every touch of VWAP as an automatic bounce, rather than a level that sometimes gets sliced straight through on strong momentum.
• Forgetting that VWAP is backward-looking by construction: it's an average of everything that has already traded, not a projection of where price is headed.
• Ignoring anchored VWAP entirely and only ever using the default session reset, missing the chance to judge a specific catalyst's staying power.

Watching This for Free

VWAP is a free, standard indicator on TradingView, most retail broker platforms, and day-trading-focused charting tools, typically with both session and anchored variants available at no cost.

• 1Add VWAP to an intraday chart on any free charting platform; it's usually already available as a one-click session overlay.
• 2Watch whether price spends most of the session above or below the line as a simple read on which side is in control.
• 3Look for a break-and-retest of VWAP as one possible confirmation signal, alongside a momentum indicator from earlier in this module.
• 4Try anchoring VWAP to a specific catalyst (an earnings date or a major gap) instead of the default session reset.
• 5Remember VWAP is a benchmark of what already happened, not a prediction. Pair it with price action before acting on it.

Try It Yourself: Compare Session vs. Anchored VWAP

1. 1Pull up any liquid stock on an intraday chart on a free charting platform and add the standard session VWAP.
2. 2Note how many times price crossed VWAP during the session, and whether those crosses led anywhere or reversed immediately.
3. 3Find a recent earnings date or major gap for the same stock and add an anchored VWAP starting from that point.
4. 4Compare where price sits relative to the anchored VWAP today versus where it sat on the anchor day itself.
5. 5Decide, honestly, whether that anchored VWAP tells you anything the session VWAP alone couldn't.

VWAP closes out the volume-based half of this module's toolkit alongside OBV and ATR: none of the three care about direction the way RSI, MACD, or the moving averages do. Instead they measure conviction, magnitude, and fair value: three different lenses on the same underlying question of whether a price move is actually backed by real participation.

Visualize It: Build a VWAP From Three Prints

VWAP weights each price by the volume traded there. Set three price levels with their volumes and compare with a plain average.

No Inventor: Just a Trading Desk's Benchmark

VWAP doesn't have a named creator the way most indicators in this module do. It emerged from institutional trading desks in the 1980s as electronic execution started replacing floor trading, as a simple answer to a practical question: did this trade get filled at a fair average price relative to the rest of the day's volume? Asset managers and brokers adopted it as a shared benchmark for execution quality. A fund manager could prove to a client that a large order was filled near VWAP, not dumped carelessly into the market.

Cypher Guide: VWAP From Typical Price, With a Spread Band and an Anchor

This guide fixes the vocabulary of VWAP and then builds the line from raw bars: typical price, the two running sums, the quotient, and the weight each bar carries. A spread band around the line and an anchored version are worked in the same way. All figures are hypothetical.

Mechanism one: the running quotient. The first three bars reproduce the chapter: typical prices of 100.00, 101.00 and 102.00 with volumes of 10,000, 8,000 and 12,000 give a VWAP of 101.07. Two more bars are added, one quiet and one heavy. Each bar is built from a high, a low and a close, so the typical price is shown rather than assumed.

Read across the row for bar 5. The typical price fell to 100.50 on 20,000 shares, the largest bar so far. VWAP dropped from 101.14 to 100.91, a fall of 0.23, while the plain average fell only to 101.00. A heavy bar moves the line much more than a quiet one: bar 5 holds 35.7 percent of the weight. If bar 3 had traded 24,000 shares instead of 12,000, VWAP after three bars would be 101.33 rather than 101.07.

Mechanism two: the anchor. Session VWAP uses all five bars. An anchored VWAP started at bar 3, the high-volume rally bar, sums only bars 3, 4 and 5. The two lines answer different questions: where the day’s average cost sits, and where the average cost since that event sits.

Read across the anchored row. Dropping the first two bars lifts the line to 101.13 because those bars traded at lower prices. Neither value is better. Each is a different average of the same trades, and a price below the anchored line means the average buyer since the anchor is behind.

In plain words: VWAP is total money traded divided by total shares traded, so heavy bars set it and the starting point decides what it describes. Education, not advice.

Closing Rundown: What VWAP Averages and Why the Anchor Matters

Chapter 8 set out to show how VWAP is built from price and volume and why it is a benchmark first and a chart line second. The one idea is that VWAP is total money traded divided by total shares traded, so heavy bars set it and the starting point decides what it describes.

• Weight by volume. VWAP = cumulative (typical price × volume) ÷ cumulative volume. For bars at 100.00, 101.00 and 102.00 on 10,000, 8,000 and 12,000 shares it was 101.07, above the simple average of 101.00 because the last bar was heaviest.
• Heavy bars move it most. With only 2,000 shares on the third bar the same prices give 100.60. In the five-bar guide, a single bar holding 35.7 percent of the volume pulled the line from 101.14 to 100.91.
• It grades fills. A buyer who filled at 101.30 against a VWAP of 101.07 had slippage of 0.23. Execution desks are judged on that gap, and the idea dates from a 1984 order.
• Charts borrow it as a level. Price above VWAP reads as net buying control and price below as net selling control. A break, retest and hold is the common setup, never a guarantee.
• The session line resets. Standard VWAP restarts every morning, so it describes one day only. An anchored VWAP starts at a chosen event and asks whether the average buyer since then is ahead or behind.
• It is a benchmark, not a forecast. It reports how the stock traded on average. It is combined with other confirmation and not used as a standalone signal.

Chapter 9: The Stochastic Oscillator — Where StochRSI Actually Came From

Chapter 2 of this module covered StochRSI, the stochastic formula applied to RSI instead of price. This chapter covers the original: the plain Stochastic Oscillator, applied directly to price itself, decades before anyone thought to run it through RSI first. If StochRSI felt twitchier and faster than RSI, that's because it inherited its personality from this indicator.

Streamline View: How the Stochastic Oscillator Is Built and Read

The Stochastic Oscillator shows where the latest close sits between the highest high and lowest low of a recent window, as a percentage. The rule that matters most is Lane’s observation that momentum turns before price does: as a trend tires, closes drift away from the extremes of the range before price reverses. The usual ways it goes wrong are selling every touch of 80 and buying every touch of 20 in a strong trend, and trading every %K and %D cross in a market that is not ranging.

Second outcome. A close of 146.00 in the same window gives %K of 5.0, deep in the oversold zone, and a close at 165.00 gives 100.0. Now suppose the window’s high rises to 170.00 on a later day while the close stays at 159.00. %K falls to 56.0 even though price did not drop, because the range grew. In a strong uptrend the line can stay above 80 for a long time, which is why an extreme reading is a prompt to look at the trend and not a sell order.

In plain words: the Stochastic Oscillator says where the close sits in its recent range, its crossovers are frequent, and Lane himself paired it with other tools. Education, not advice.

History of the Stochastic Oscillator: A Classroom Method

The Stochastic Oscillator is one of the oldest oscillators still on every chart, and the one that supplied the %K and %D vocabulary used across this module. Its history is mainly a story of a classroom: a trader who taught for decades, wrote it up in a magazine in 1984, and saw its formula reused on other indicators long after.

• The author never meant it to stand alone. Lane taught it with other tools, among them Elliott Wave and Fibonacci retracements, for timing. A trader who uses only the 80 and 20 lines is using a smaller version of the method than the one he taught.
• The core claim is about order, not level. Lane is quoted as saying that momentum changes direction before price does. The claim concerns when the line turns relative to price, which is why divergence is the classic reading and a bare zone touch is the weak one.
• It is the parent of a family. Because the formula is a position-in-range test, it was later pointed at RSI and at other series. A reading means something different depending on whether the input is raw price, as here, or an already smoothed indicator.

Who Built This, and Why

George Lane developed the Stochastic Oscillator in the late 1950s, while working as a securities trader and educator at E.F. Hutton & Co., where, notably, he trained under Joseph Granville, the same person who would later invent On-Balance Volume, covered in Chapter 6. By Lane's own account, his team tested 28 different oscillator designs before settling on this one and expressing it as a percentage. His central idea was that momentum turns before price does: as a trend loses steam, closing prices start drifting away from the extremes of the recent range well before the price itself actually reverses.

The Formula: Where Today's Close Sits Inside the Recent Range

%K measures exactly one thing: where the most recent close sits, as a percentage, between the highest high and lowest low of the lookback window (14 periods by default):

A close sitting right at the top of the 14-period range prints a %K near 100; a close sitting right at the bottom prints near 0. %D is just a smoothed 3-period average of %K, and the two lines crossing each other is the indicator's most-watched signal, the same %K/%D crossover mechanic Chapter 2 applied to RSI instead of raw price.

Reading %K and %D: Overbought, Oversold, and the Crossover

The standard thresholds are 80 for overbought and 20 for oversold, with 50 acting as a rough centerline between the upper and lower half of the range. A %K line crossing above %D from below is read as a bullish signal; crossing below from above is read as bearish, though standard charting documentation is blunt that these crosses are "frequent and prone to whipsaw," especially outside of a clearly range-bound market. Bullish divergence (price making a lower low while the oscillator makes a higher low) and bearish divergence (price making a higher high while the oscillator makes a lower high) are watched the same way they are for every other oscillator in this module.

A Real Example: Revisiting the Fastest RSI Sprint Since 1982

Chapter 1 covered a real, dated event: the S&P 500's 14-day RSI went from oversold (below 30) to overbought (above 70) in just 12 trading days following the March 30, 2026 low, the fastest such move since 1982, according to reporting at the time. RSI measures average gains versus average losses; the plain Stochastic Oscillator measures something related but distinct: where the close sits within the recent high-low range. Because raw price %K tends to be even more reactive to short-term extremes than RSI's smoothed average-gain/average-loss calculation, a plain stochastic reading on the S&P 500 during that same 12-day sprint would very plausibly have touched its own 80 overbought threshold even earlier than RSI crossed 70 — the entire reason George Lane's original design was later run through RSI to build StochRSI in Chapter 2, and it's the same underlying dynamic showing up a third way.

How Professionals Actually Use It

The Stochastic Oscillator is most trusted in genuinely range-bound markets, where its overbought and oversold readings tend to mark real turning points rather than getting steamrolled by a strong trend. In trending conditions, desks tend to watch the 50 centerline more than the 80/20 extremes, treating %K holding above 50 as confirmation a pullback is still just a pullback within an uptrend, not the start of a reversal. Divergence between %K and price, especially on longer timeframes, is weighted more heavily than any single crossover.

Common Mistakes

• Selling every touch of 80 or buying every touch of 20 without checking whether the broader trend is strong enough to keep the oscillator pinned at an extreme.
• Trading every %K/%D crossover in a trending market, where standard charting documentation itself warns these crosses are frequent and prone to whipsaw.
• Confusing this indicator with StochRSI from Chapter 2: they share a formula shape but operate on entirely different inputs (raw price here, RSI there).
• Ignoring George Lane's own stated approach of combining it with other tools, and trading it as a complete standalone system instead.

Watching This for Free

The Stochastic Oscillator is a standard, free indicator on TradingView, Yahoo Finance, and virtually every retail broker's charting platform, typically listed right next to RSI and StochRSI.

• 1Add the Stochastic Oscillator to any free charting platform, using the 14/3/3 default settings.
• 2Identify whether the chart is currently range-bound or trending. That context changes how much weight to give an 80/20 reading.
• 3Watch %K/%D crossovers near the 50 centerline as a lower-drama confirmation signal than crosses at the extremes.
• 4Compare this chapter's plain price-based reading against StochRSI from Chapter 2 on the same chart to see how differently the two behave.
• 5Pair any signal with a second tool from earlier in this module, following Lane's own stated approach rather than trading the oscillator alone.

Try It Yourself: Compare Stochastic to StochRSI

1. 1Pull up any liquid stock on a free charting platform and add both the plain Stochastic Oscillator and StochRSI from Chapter 2.
2. 2Find a recent sharp price move and note which of the two touched its overbought or oversold threshold first.
3. 3Count how many %K/%D crossovers each indicator produced over the last month, and note which one whipsawed more.
4. 4Check whether either indicator's divergence from price lined up with an actual reversal, or continued straight through it.
5. 5Decide, honestly, which version felt more useful on this particular chart, and note that the answer will likely change on a different stock or timeframe.

This chapter closes a loop that opened all the way back in Chapter 2: StochRSI wasn't invented from nothing, it was built by taking this exact formula and pointing it at RSI instead of price. Seeing the original next to its descendant is a useful reminder that most "new" indicators in this field are really just familiar math applied to a different input.

Visualize It: Read a Stochastic From the Range

The Stochastic compares the latest close with the high-low range of the window. Set the range and the close.

The Educator Who Said Momentum Turns Before Price Does

George Lane ran Investment Educators Inc. starting in the 1950s, training brokers and traders rather than managing money himself. His core observation was that as a trend loses strength, the closing price tends to drift away from the high (in an uptrend) or the low (in a downtrend) before the price itself actually reverses — momentum peaks early. The Stochastic Oscillator was his way of quantifying that drift. Lane taught it in seminars for decades before it became standard software output, which is part of why the %K/%D naming still carries his classroom shorthand.

Cypher Guide: %K and %D Rebuilt, Fast Against Slow, and Window Exit

This guide fixes the vocabulary of the Stochastic Oscillator and then works the calculation by hand: the chapter’s %K and %D rebuilt from closes, the fast and slow versions side by side, and what happens when an old extreme leaves the window. All figures are hypothetical.

Mechanism one: rebuilding the chapter. The chapter gives a lowest low of 145.00, a highest high of 165.00 and a close of 159.00, so %K is (159 − 145) ÷ 20 = 70. Its three-day %K values of 62, 68 and 70 correspond to closes of 157.40, 158.60 and 159.00 if the range stayed fixed. The inverse is useful: close = lowest low + %K ÷ 100 × range.

Read across the formula lines. One point of %K is worth 0.20 of price here, so the move from 62 to 70 is 1.60 in price. The indicator turns a small price change into a large scale change whenever the range is narrow.

Mechanism two: fast against slow. The fast version uses the raw value as %K and its 3-bar average as %D. The slow version first smooths raw %K, then averages the result again. The table uses six raw values, the chapter’s 62, 68 and 70 padded by 55 before and 64 and 58 after.

Read across day 5. Raw %K has fallen from 70 to 64, a drop of 6, while the slow %K is 67.3 and the slow %D is 65.2. Fast %D equals slow %K by construction. The slow signal line first exists on day 5, two bars after the fast one, so the slow pair answers later and crosses less often.

Window exit. Suppose the 165.00 high was set 14 bars ago and now leaves the window, and the next highest high is 161.00. The close of 159.00 has not moved.

Read across the row for a range of 60. The same 14.00 above the low reads only 23.3. The oscillator measures position inside the range, so it does not say how large the move was.

In plain words: the Stochastic Oscillator reports where the close sits between the window low and high, so it depends on the window as much as on price. Education, not advice.

Closing Rundown: What the Stochastic Oscillator Reads From the Range

Chapter 9 set out to show how the Stochastic Oscillator reports where the latest close sits inside the recent high-low range, and how %K and %D are read. The one idea is that the reading depends on the window as much as on price, and an extreme reading is a prompt to check the trend and not a sell order.

• Position in a range. %K = (close − lowest low) ÷ (highest high − lowest low) × 100. With a low of 145.00, a high of 165.00 and a close of 159.00, %K is 70.0 and %D, the average of 62.0, 68.0 and 70.0, is 66.7.
• The range can move the reading. If the window high rises to 170.00 with the close fixed at 159.00, %K falls from 70.0 to 56.0. If the 165.00 high leaves the window and the next high is 161.00, it rises to 87.5.
• Narrow ranges magnify. The same 14.00 above the low reads 70 in a range of 20 but only 23.3 in a range of 60. The reading does not say how large the move was.
• Fast and slow answer differently. The slow version smooths %K again, so its signal line first exists two bars later and crosses less often. Fast %D equals slow %K by construction.
• Crosses are frequent. A %K cross above %D reads bullish and below reads bearish, with 80 and 20 as the usual lines. They whipsaw outside a ranging market, so a reading of 70 is not yet overbought.
• Use it with other tools. Its originator paired it with other methods, and the 2026 rally continued well past the first overbought flags. A touch of 80 in a strong trend is not a sell signal.

Chapter 10: Average Directional Index (ADX) — Measuring Trend Strength, Not Direction

This is the third indicator in this module built by J. Welles Wilder (the same person behind RSI in Chapter 1 and ATR in Chapter 7) and it solves a problem neither of those tools was designed for. RSI and the moving averages can tell you which way price is leaning. ADX answers a completely different question: is there actually a real, sustained trend here at all, or is this just noise?

Streamline View: How ADX Is Built and Read

The Average Directional Index (ADX) measures how strong a trend is, whichever way it points. Two companion lines, +DI and −DI, give the direction. The rule that matters most is to keep the two jobs apart: ADX rising means the trend is gaining strength, not that price is going up. The usual ways it goes wrong are reading a rising ADX as bullish, and expecting it to react quickly, since Wilder’s heavy smoothing makes it slow.

Second outcome. Swap the two lines, +DI 14.2 and −DI 32.5, and DX is the same 39.2, so ADX would rise the same way, but the trend is downward. With +DI at 22 and −DI at 20, DX is only 4.8: the lines are nearly tied and there is little trend to measure. The same ADX value can sit under very different pictures, so read the DI lines first.

In plain words: ADX says how strong a trend is and never which way it points, the DI lines give direction, and a slow ADX is the price of smoothing out noise. Education, not advice.

History and Origins of ADX

ADX has a short public record. It was published once, in 1978, by an engineer who built it as a chain of explicit steps, and its formulas have not changed since. What changes from chart to chart is the smoothing a platform uses, not the idea. This section therefore looks at the pieces it was built from and what each of them brought.

• It was built to rank markets. ADX began as a screening tool for commodities, a way to ask which market is trending hardest. That is why it says nothing about direction by itself, and why professionals use it as a filter before choosing a method. The same logic applies to a single chart: a low reading does not mean the market is about to fall, only that there is no strong trend to follow right now.
• The thresholds are conventions. Levels such as 20 and 25 are the ones in common use. They are not laws of the formula, and many analysts use 20 for more sensitivity. Treat them as a working guide and look at the shape of the line.
• The smoothing choice changes the numbers. Wilder smoothed with his own recursive average. Replacing it with a plain simple or exponential average gives different values, so two platforms can show different ADX readings for the same bars and still both follow his design. When comparing a chart with a book or a lesson, check which smoothing it uses, and compare like with like before concluding that a reading is high or low.

Who Built This, and Why

Wilder introduced ADX alongside RSI and ATR in the same 1978 book, New Concepts in Technical Trading Systems, originally built for commodity markets. His insight was that direction and strength are two separate measurements that most indicators conflate: a stock can be technically "trending up" by the thinnest possible margin, or trending up with real, measurable conviction, and traditional tools didn't distinguish between the two. ADX was designed specifically to fill that gap.

The Formula: Direction and Strength, Calculated Separately

The system starts by comparing each day's high and low to the day before's, to see which direction had the bigger move:

Those raw directional movements get smoothed over 14 periods and divided by ATR from Chapter 7 to produce +DI and −DI, the actual directional indicators. The gap between them, expressed as a percentage of their sum, becomes DX, and ADX itself is just a smoothed moving average of DX:

Reading ADX: The Threshold Levels That Actually Matter

Wilder's original guideline treats ADX below 20 as "no meaningful trend," 20–25 as a gray zone, and above 25 as a genuinely strong trend, though charting documentation notes many modern analysts use 20 as the key line instead, favoring more sensitivity. Critically, ADX rising says nothing about direction on its own; a rising ADX during a downtrend means the downtrend is strengthening,

not that a reversal is coming. Direction still has to come from +DI and −DI: a common signal is +DI crossing above −DI while ADX sits above the strength threshold, read as a trend-confirmed buy signal, with the mirror image read as a trend-confirmed sell.

A Real, Dated Example: What ADX Was Built to Flag

Nvidia's 2023–2024 run is one of the clearest real illustrations of what a genuinely strong, ADX-confirmed trend looks like, even though the specific ADX readings from that period aren't something financial media typically quotes the way they quote a price or a percentage gain. The stock rose 239% across 2023 alone, roughly ten times the S&P 500's return that year. It kept climbing into 2024, gaining 36% just through mid-February, then jumping 16.4% the day after its February 21, 2024 earnings report and 17.2% over the following three sessions, on revenue growth of 265% year-over-year. A second earnings report on May 22, 2024 added another 9.3% same-day and 20% over the next three sessions. By the first half of 2024 alone, the stock had climbed 150%.

How Professionals Actually Use It

ADX is rarely traded as a standalone entry signal: it's used as a filter that decides which strategy applies. When ADX is low, desks favor range-trading approaches, like the Bollinger Band and RSI mean-reversion setups covered earlier in this module. When ADX climbs above the strength threshold, the same desks shift toward trend-following approaches, since mean-reversion tools tend to fail badly in a genuinely strong trend. Some traders also watch ADX's own slope — a rising ADX from a low level is read as a new trend just getting started, while a falling ADX from a high level is read as an established trend losing steam, even before +DI and −DI actually cross.

Common Mistakes

• Reading a rising ADX as automatically bullish, when it only measures strength: a rising ADX in a downtrend means the downtrend is getting stronger, not reversing.
• Trading every +DI/−DI crossover regardless of the ADX level, when standard charting documentation itself warns these crosses need filtering with complementary analysis.
• Using mean-reversion strategies like Bollinger Band fades while ADX is elevated, fighting a trend the indicator is explicitly telling you is strong.
• Expecting fast signals: Wilder's heavy smoothing means ADX needs a substantial amount of price history before its readings become reliable.

Watching This for Free

ADX, +DI, and −DI are standard, free indicators on TradingView, Yahoo Finance, and most retail broker platforms, typically bundled together as a single "ADX" indicator package.

• 1Add ADX with +DI and −DI to any free charting platform, using the 14-period default.
• 2Check the current ADX level first: below 20 favors range strategies, above 25 favors trend-following ones.
• 3Use +DI and −DI crossovers for direction only once ADX confirms there's a real trend to trade.
• 4Watch ADX's own slope: rising from a low level suggests a fresh trend building, falling from a high level suggests fading momentum.
• 5Switch strategies as ADX crosses the threshold rather than forcing the same approach in every market condition.

Try It Yourself: Find a Real Strong Trend

1. 1Pull up a stock you know had a sustained, powerful run (Nvidia's 2023–2024 move is a well-documented real example) and add ADX on a free charting platform.
2. 2Note where ADX sat during the strongest part of that run compared to a calmer, more range-bound period for the same stock.
3. 3Find a period where +DI and −DI crossed while ADX was still below 20, and check whether that signal held up or whipsawed.
4. 4Compare that to a crossover that happened while ADX was above 25, and note any difference in how the trade would have played out.
5. 5Decide, honestly, whether checking ADX first would have changed which setups you'd have taken on that chart.

Between RSI, ATR, and now ADX, this module has covered three separate tools from the same 1978 book, each one measuring something completely different from the same underlying price data. That's a useful reminder that "technical analysis" isn't one method; it's a toolbox, and even a single analyst's own body of work can cover direction, magnitude, and strength as three entirely separate questions.

A Third Tool From the Same Single Book

ADX is the third Wilder indicator in this module, from the same 1978 book as RSI and ATR. Wilder built it to answer a question the other two couldn't: not whether price was overbought or volatile, but whether a market was trending at all versus just chopping sideways. The +DI/-DI/ADX system was deliberately designed to be read together: ADX for trend strength, the DI lines for direction, which is why it's usually taught as one combined tool rather than three separate ones.

Cypher Guide: ADX From Raw Bars, Directional Movement to the Smoothed Index

This guide fixes the vocabulary of ADX and then works the calculation by hand: how a bar is classified into directional movement, how the smoothed sums become +DI, −DI and DX, and how the final smoothing slows ADX down. A short period of 7 is used for the first example so the arithmetic fits on a page; the standard period is 14. All figures are hypothetical.

Mechanism one: classify, sum, divide. The table lists nine bars as high, low and close. For each bar from 1 to 8 it finds UpMove, DownMove, the two DMs and the true range. Bars 3 and 5 are inside bars (both moves are zero or negative) and bar 4 is an outside bar. The first seven changes are then summed.

Read across bars 3 to 5. The two inside bars each contribute 0 to both columns, yet their true ranges (3 and 4) still go into the denominator. Bar 4, an outside bar, has UpMove 3 and DownMove 2, so only +DM = 3 counts. After bar 8 the lead flips: +DI falls from 20.6 to 17.6 while −DI rises from 14.7 to 18.4. The flip comes after only one more down bar because the window is 7 bars long; with 14 the sums would respond more slowly.

Mechanism two: the second smoothing. DX can swing from bar to bar, so ADX averages it. The chapter’s numbers give +DI = 32.5 and −DI = 14.2, so DX = 39.2. Suppose the previous ADX was 24.

Read across the table. Even with DX held at 39.2, ADX takes about six bars to climb from 18 to 25. That delay is why a trend is often well under way before ADX confirms it. For the same reason ADX can keep rising for a few bars after DX has begun to fall, since it is an average of past values.

In plain words: +DI and −DI split the true range into upward and downward share, DX measures how lopsided they are, and ADX averages that lopsidedness slowly, so it confirms a trend late and says nothing about direction. Education, not advice.

Closing Rundown: What ADX Measures and What It Leaves to the DI Lines

Chapter 10 set out to show how ADX measures the strength of a trend and how the two DI lines supply its direction. The one idea is to keep those two jobs apart: a rising ADX means the trend is strengthening, not that price is rising.

• Two jobs, three lines. +DI and −DI give direction, while ADX gives strength. With +DI at 32.5 and −DI at 14.2, DX = 100 × |32.5 − 14.2| ÷ (32.5 + 14.2) = 39.2.
• Classify each bar. UpMove is today’s high minus yesterday’s high and DownMove is yesterday’s low minus today’s low. Only the larger positive one counts as directional movement, so inside bars add zero.
• ADX is heavily smoothed. ADX = (previous ADX × 13 + DX) ÷ 14. A DX of 39.2 moved ADX only from 22.0 to 23.2, and with DX held at 39.2 it still took about six bars to climb from 18 to 25.
• Swapping the DI lines changes direction, not strength. With +DI 14.2 and −DI 32.5 the DX is the same 39.2, but the trend points down. Read the DI lines first.
• Thresholds are guidelines. Below 20 means no meaningful trend, 20 to 25 is a gray zone and above 25 is a strong trend. Many analysts use 20 as the key line for more sensitivity.
• It confirms late. Because ADX averages past values, a trend is often well under way before it confirms and it can keep rising briefly after DX turns down.