Seeing Patterns

Learning to See What the Chart Is Telling You

A pattern is an indicator, although it is an unusual kind of indicator because it has traditionally been drawn directly on the chart by the trader. Today, charting software can identify and draw many of these formations for you, and some artificial-intelligence tools can do the same thing. That convenience does not make pattern recognition objective, however. A pattern still depends upon the prices you choose to connect and the formation you believe those prices are creating. You may prefer the apparent precision of a mathematical indicator, but remember that many successful traders have built their methods around patterns alone. There is another reason patterns deserve your attention. The most widely recognized formations are no longer secrets. They appear in trading books, financial websites, television programs, and market commentary, which means that thousands of traders may be watching the same formation and reacting to it at approximately the same time. When enough people recognize the same pattern, their actions can help produce the very price movement the pattern is supposed to forecast.

Technical analysts have been looking for recurring formations since the earliest days of chart analysis. The underlying idea is fairly simple. A market can make a substantial move in one direction, pause or retreat as traders take profits and reconsider their positions, and then resume the original move. That broad sequence occurs repeatedly. From it came the familiar formations you see on charting sites today: head-and-shoulders patterns, double tops and bottoms, cup-and-handle formations, flags and pennants, and ascending and descending triangles, among many others. Do not make the mistake of assuming that one list contains all the patterns worth knowing. A pattern becomes useful because it repeatedly appears and behaves in a particular way on the securities you trade. Some stocks, indexes, or commodities develop certain formations again and again, and traders who follow those markets closely may recognize the behavior before everyone else does.

You have already encountered several patterns elsewhere in this book. The inside day belongs in this category, as do support and resistance lines, island reversals, and the various candlestick formations. Fibonacci retracements are another example. Traders have created hundreds of named formations over the years, and there is little value in trying to memorize every one of them. Learn a manageable group and learn them well. Knowing five patterns thoroughly is far more useful than recognizing fifty names without understanding how the formations behave.

Introducing Patterns

Chart patterns are geometric formations that appear when you connect important price points on a chart. A triangle is the obvious example, but patterns can also form rounded structures such as a cup and handle or a head-and-shoulders formation. Like other technical indicators, a pattern is not merely a description of what has already happened. The formation carries an expectation about what prices may do next. Most patterns use straight trendlines, although some depend more heavily on curved or rounded price action. Those lines generally connect a sequence of highs, lows, or both. Once the formation has been identified, it is classified according to what it is expected to do: continue the existing trend or reverse it.

The terminology can become almost comical. Traders talk about double bottoms, dead-cat bounces, belt holds, scoops, fry-pans, cradles, and jay-hooks as though they were describing objects in a hardware store. Some names sound ridiculous until you see the formation they describe. Then the name often makes the pattern easier to remember. Modern traders also have an advantage that earlier chartists did not. You can watch analysts demonstrate formations on trading websites and video platforms, compare examples, and return to the material repeatedly until the shapes become familiar. Pattern references and historical studies can also give you a statistical framework for judging how formations have behaved in the past. Historical performance is not a promise about the next trade, but it is much better than pretending every pattern works equally well.

A developing pattern is never guaranteed to become the pattern you think you see. This is one of the first frustrations you encounter when learning chart formations. The price may appear to be building a triangle, for example, and then suddenly break through the line you were watching and destroy the formation. A possible head-and-shoulders pattern can become something else entirely. You will make these mistakes. Expect them. Pattern recognition is partly an exercise in waiting for enough evidence to accumulate before you give the formation a name. The reason you put up with this uncertainty is that a completed and properly identified pattern can provide a useful forecasting framework. When the pieces finally fit together, the chart can tell you considerably more than a collection of individual bars.

Using Your Imagination

Humans are exceptionally good at seeing shapes and relationships. The difficulty is that we are also exceptionally good at seeing shapes that do not really exist. Put a month of price data in front of a beginner and he may see nothing but a meaningless collection of bars. Give him some experience and suddenly triangles, channels, tops, bottoms, and other formations begin appearing everywhere. This is both the advantage and the danger of pattern recognition. You need enough imagination to notice a developing formation, but enough discipline to reject formations that do not satisfy the basic requirements of the pattern.

At first glance, a raw price chart may simply look like a series of prices drifting around without any obvious organization. Look closer, and a subtle structural pattern may be taking shape. When price action begins making a sequence of progressively lower highs while simultaneously forming a series of higher lows, it creates a symmetrical triangle. Connecting those lower highs with an upper trendline and the higher lows with a lower trendline reveals two converging lines that gradually point toward an apex. The price cannot remain inside this narrowing range forever; as trading continues, price action will eventually be forced to break through one of the two converging boundaries.

The interesting question is not whether the triangle will eventually break. The interesting question is which side will break first. Because the majority of the bars in this example are moving downward, you might reasonably expect the eventual breakout to occur through the lower trendline. Historical studies of symmetrical triangles have shown that downside breakouts occur more than half the time, which gives the formation a continuation tendency in this particular context. But that tendency is not a guarantee. Upside breakouts occur as well, and the same historical work found them occurring a substantial minority of the time. When the upside breakout happens, the triangle functions as a reversal rather than a continuation pattern.

Notice something important here. A pattern does not always hand you a buy or sell signal while it is still being constructed. During much of the life of a symmetrical triangle, all you really know is that the range is narrowing and that a breakout will eventually be required. You may have a directional suspicion, but you do not yet have confirmation. The temptation is to act early because you believe you know what is coming. Resist that temptation when the pattern rules do not yet give you enough evidence.

Volume often adds another piece to the puzzle. As a formation approaches completion, trading volume frequently increases as more market participants recognize what is taking shape. They are not necessarily following the same reasoning you are, but they are responding to the same visible price structure. Triangles are particularly interesting because volume commonly contracts while the formation develops and then expands around the breakout. That shrinking activity can itself be a warning. The market is becoming compressed, and compressed price action eventually has to resolve itself.

Staying Inside the Lines

Do not expect every price point in a pattern to behave with geometric perfection. Real markets are not drawing-board exercises. If you are constructing a resistance line from several highs, for instance, you should not throw away the entire pattern because one bar extends slightly above the line. What matters is whether enough significant highs or lows establish the boundary clearly enough for the line to have meaning. Triangles, flags, and pennants all depend on this kind of support and resistance structure, and a certain amount of imperfection is normal.

There is disagreement among technicians about exactly how much imperfection should be tolerated. Some analysts permit one or two bars to penetrate a trendline slightly, particularly when the penetration is small and price quickly returns inside the formation. Other technicians take a much stricter position and argue that once the trendline has been violated, you should treat the violation as meaningful rather than explain it away. The practical lesson is that you need rules before you draw the pattern. Otherwise, you will find yourself changing the rules after the fact simply to preserve a formation you want to believe in.

You also do not have to depend entirely on your eyes. Pattern recognition can be strengthened by combining the formation with other indicators. Moving averages, for example, can provide confirmation of the direction suggested by a triangle or another pattern. The purpose is not to make the pattern prettier. The purpose is to have another piece of evidence supporting the interpretation before committing capital.

Getting Comfortable With Continuation Patterns

A continuation pattern develops when the dominant buying or selling pressure temporarily loses momentum without actually disappearing. The market pauses. Traders take profits, new participants enter, and those already holding positions reconsider what they should do next. If the larger trend remains intact, the pause can eventually resolve in the same direction as the move that came before it. This makes continuation formations particularly interesting when you already hold a position and are looking for an opportunity to add to it.

Continuation patterns are frequently short-lived. Some may last only a handful of trading sessions, which means they are easy to overlook if you are not watching the chart closely. Their small size can also make them less impressive than major reversal formations. Do not confuse size with importance. A short pause in a powerful trend can provide an opportunity to participate in another leg of the move, provided the evidence supports continuation.

Ascending and Descending Triangles

The construction of an ascending or descending triangle begins much like the symmetrical triangle. You identify a series of highs and lows and draw the appropriate boundaries. The difference is that one boundary becomes horizontal while the other slopes. The resulting shape tells you something about the balance between buying and selling pressure.

An ascending triangle contains a relatively flat series of highs forming a horizontal resistance level, while the lows continue to rise beneath it. At first, the inability to push above the old highs may look bearish. After all, the market keeps running into the same ceiling. But look at what is happening underneath. Buyers are stepping in at progressively higher prices. The market is refusing to make new lows, and that rising line of lows indicates that demand is gradually becoming more aggressive. Eventually the resistance line may give way and produce an upside breakout.

Whenever you identify a horizontal series of highs, look immediately beneath it for rising lows. The upward-sloping support line does more than complete the triangle. It also gives you a logical reference point for risk management because a decisive move below that support can tell you that the expected continuation is no longer behaving as anticipated. Historical studies have found that ascending triangles produce the expected upward move roughly two-thirds of the time, while the remainder fail to produce the anticipated result. Waiting for the price to actually close above the upper trendline can greatly reduce the historical failure rate reported in those studies. The expected move after the breakout is commonly estimated from the height of the triangle, a technique known as the measured move.

The descending triangle reverses the arrangement. Here the lows tend to hold near a horizontal support level while the highs continue to decline. The prevailing trend is downward, but the market appears unable to make fresh lows. That may tempt you to conclude that the decline has finally exhausted itself. Look at the falling highs before making that assumption. Sellers are still willing to enter at progressively lower prices, and that descending resistance line tells you that downward pressure remains. Buying simply because the market has stopped making new lows can therefore be premature. The pattern still favors continuation of the decline until the evidence says otherwise.

The Dead-Cat Bounce

A dead-cat bounce is one of the stranger continuation formations because it initially looks like the beginning of a reversal. The market has been falling, suddenly stages a substantial upward retracement, and gives traders the impression that the decline is finished. Then the bounce loses its strength and the original downward trend resumes. The formation is most often associated with falling markets, although an upside version can occur as well.

The name is memorable because the move is deceptive. The rally can be strong enough to attract buyers who believe the worst has passed. In some cases, the recovery even retraces far enough to fill a previous price gap. That can be particularly misleading because traders often assume that filling a gap means the move that created it has ended. It does not. A gap can be filled while the original trend remains completely intact. Historical studies of these formations have found that only a little more than half of the associated gaps had closed within six months, illustrating why a gap fill should not automatically be treated as proof of a trend reversal.

The dead-cat bounce has also shown a high historical success rate in long-term pattern studies. That does not mean every sharp rally during a decline is a dead-cat bounce. The surrounding trend, the size of the retracement, the volume behavior, and what happens after the bounce all matter. The mistake is to see the rally and immediately declare the bear move over. The market has fooled traders this way many times before.

Cup and Handle

The cup-and-handle formation was popularized by William J. O’Neil and became one of the better-known bullish chart patterns. It is called a cup because the first part of the formation resembles a rounded bottom. The handle develops afterward as the price pulls back or moves sideways in a relatively narrow downward-sloping channel. The complete formation can require a great deal of time to develop, sometimes many months or even more than a year, which is one reason you should not expect to find a valid example every time you look at a chart.

The formation begins with a decline that gradually rounds into a bottom and then recovers toward the level where the earlier decline began. The recovery does not necessarily stop at exactly the old high. It may reach somewhat above or below it before another pullback begins. That pullback forms the handle. This is the part of the pattern that can test your patience because it may look as though the attempted recovery has failed. In a properly developing cup and handle, however, the important event is what happens next. You are watching for the price to break above the upper boundary of the handle and resume the advance.

The cup and handle can be difficult to distinguish from ordinary price fluctuations, particularly when you look at a formation before it has completed. That is why confirmation from other indicators can be useful. The pattern gives you a framework for interpreting the price action, but it does not remove uncertainty. As with every pattern in technical analysis, the formation is a probability statement about what may happen next, not a promise that the market will obey the shape you have drawn on the chart.

Recognizing Classic Reversal Patterns

Patterns become particularly valuable when they help you recognize that an established trend may be running out of steam. A market does not simply announce, “The trend is over.” Instead, the change usually develops through a sequence of price movements, and those movements can leave behind recognizable formations. No reversal pattern guarantees that a trend will change, but certain formations have become important enough that you should know what they look like and, more importantly, what must happen before you treat them as valid.

Double Bottom

A double bottom has the appearance of the letter W. The market declines to a low, rallies, returns to approximately the same low, and then attempts another advance. The important point is that the second visit to the low represents a test of the first decline. If sellers cannot push the price materially below the original low and buyers subsequently take control, the formation can become a bullish reversal pattern. But you do not have a valid double bottom merely because a chart happens to contain two lows.

There are several characteristics that help distinguish a genuine double bottom from an ordinary pair of price dips. The two lows should normally be separated by at least ten trading days, although the distance can extend to several months. The difference between the two bottoms should generally be small, with no more than about a 4 percent variation between them. Between the lows, the price should rally at least 10 percent from the lower of the two bottoms. Most importantly, the price must eventually rise above the confirmation level before the formation is considered complete.

The confirmation level is drawn horizontally through the highest point between the two lows. Think of the W as having a central peak between its two legs. That peak becomes the reference point. When price finally moves above it, the double bottom has received its confirmation. This is the event that changes the pattern from an interesting possibility into a tradable formation. Until then, you are looking at two lows and a rally between them. That is not enough.

You may also notice that individual bars occasionally wander across the lines used to outline the formation. This does not automatically invalidate the pattern. A line drawn simply to illustrate the shape of a formation is not necessarily functioning as a formal support or resistance level, so minor penetration can be tolerated in some circumstances. What matters more is whether the overall structure remains recognizable and whether the price eventually satisfies the confirmation requirement.

After the upside breakout, watch for a retracement. The price may rise above the confirmation line and then retreat toward it before beginning another advance. This behavior occurs often enough in confirmed double bottoms that you should not be surprised when it happens. Different technicians use different terminology for these retracements. A retracement following an upside breakout is sometimes called a throwback, while the same type of move following a downside breakout is often called a pullback. In ordinary conversation, however, traders use retracement, correction, throwback, and pullback with considerable freedom. The name matters less than recognizing what the price is actually doing.

Here is where the statistics become interesting. A large percentage of formations that initially resemble double bottoms never satisfy the confirmation requirement. Only about one-third of the apparent patterns in the cited studies eventually qualify, meaning that roughly two-thirds fail before confirmation. Those numbers sound discouraging until you separate an unconfirmed pattern from a confirmed one. Once price actually breaks above the confirmation line, the historical success rate changes dramatically, with the cited studies showing profitable outcomes more than 90 percent of the time. The lesson is straightforward: do not confuse the possibility of a double bottom with a confirmed double bottom.

While a classic double bottom forms a crisp, clean W, real price action is rarely so symmetrical. One or both bottoms can be rounded rather than sharp. When the first low is narrow and pointed while the second develops as a broad, rounded bottom, the formation is known as an Adam and Eve pattern. Reverse the shapes and you get Eve and Adam; make both bottoms pointed and you have Adam and Adam, while two rounded bottoms produce Eve and Eve. Technical analysis has a surprisingly imaginative vocabulary, although it rarely gets much more adventurous than this.

Time can make the formation even harder to recognize. The two bottoms may be separated by several months, and smaller retracements or additional patterns can develop inside the larger W and obscure its shape. Large formations may become easier to see on weekly charts because the longer time frame removes some of the noise present in daily price data. You may also encounter traders who identify these formations on hourly charts or other intraday time frames. There is nothing inherently wrong with that approach. The interpretation of the pattern does not magically change because the bars represent four hours instead of one day, provided the traders participating in that market are actually watching and reacting to the same time frame.

Double Tops

The double top is the opposite of the double bottom and resembles the letter M. Price reaches a high, retreats as traders take profits, and then makes another attempt to reach or exceed the previous high. When that second attempt fails, the market has demonstrated that buyers were unable to establish a new high. The balance of power may now be shifting toward sellers.

Volume can provide another clue. A genuine double top is frequently accompanied by weaker volume during the second advance. Buyers are making another attempt, but the participation is not as strong as it was during the first push upward. That does not prove a reversal, but it fits the story being told by the pattern.

As with the double bottom, you need confirmation. The confirmation level is the lowest point between the two peaks of the M. Until the price breaks below that level, you do not have confirmation that the double top has completed. Without that break, apparent double tops fail to produce a sustained decline more than half the time in the cited studies. Once the confirmation condition is satisfied, the historical results improve substantially.

There is also an important difference between tops and bottoms. Top formations frequently develop more quickly than bottom formations. Traders who have accumulated a large gain can become nervous when the market begins moving against them, and that fear can cause selling to develop rapidly. Bottoms often require more time because investors are reluctant to believe that a decline has ended and may wait for evidence before returning. For the same reason, topping formations can be more volatile.

Historical pattern studies have generally found that bottom formations outperform comparable top formations. This fits an old observation about market behavior: advancing markets often develop in a more orderly fashion than declining markets. In stocks, participation on the long side is naturally broader because anyone can buy a stock, while short selling requires a different set of conditions and is subject to additional constraints. There is also an asymmetry in the possible price movement. A stock can theoretically rise without limit, while a short position cannot profit from a decline below zero.

The Ultimate Triple Top: Head and Shoulders

Triple tops and triple bottoms are less common than their double counterparts, but the basic message is similar. The market makes repeated attempts to move beyond an important price level and fails. A head-and-shoulders formation takes this idea one step further and creates a particularly recognizable three-peak structure.

The classic head-and-shoulders pattern consists of three rallies. The first produces the left shoulder, the second rises higher and forms the head, and the third produces the right shoulder without reaching the height of the head. It is one of the most widely recognized reversal patterns in technical analysis, and the reason for that popularity is its clearly defined confirmation point. When price breaks through the neckline, the expected downward move tends to occur more than 90 percent of the time.

The neckline connects the lows separating the three major peaks. It does not have to be horizontal. In fact, a perfectly level neckline is relatively uncommon. The line may slope upward or downward, and the pattern remains a head and shoulders as long as the overall structure is intact. A downward-sloping neckline has historically been associated with a larger price decline, although the formation still requires confirmation before you treat the forecast as valid.

The psychology behind the pattern is easier to understand when you stop looking at it as three bumps and start looking at the traders creating those bumps. A head-and-shoulders generally appears after a substantial advance. The first shoulder develops when traders take profits after the market establishes another high. Buyers then return with enough strength to push the price to a new high, creating the head. The next decline brings additional profit-taking, but buyers make another attempt to rally. This time they cannot reach the previous high. The right shoulder therefore represents a failure of bullish pressure at precisely the point where buyers previously demonstrated their greatest strength.

That failure is the important part. The market has made three attempts to continue higher, yet the third attempt cannot even reach the height of the second. Once the neckline gives way, the formation has confirmed what the earlier price action was suggesting: the advance has lost its ability to maintain itself.

Do not confuse three roughly equal peaks with a head-and-shoulders pattern. If the three highs are approximately the same height, you are looking at a triple top instead. The bullish counterpart of the ordinary head and shoulders is the inverse head-and-shoulders. There you have three bottoms, with the middle bottom extending lower than the other two, and the eventual breakout is expected to occur upward.

Even after confirmation, the market may make one last trip back toward the neckline. These retracements can be frustrating because they make you question whether the breakout was real. They are often brief, lasting only a week or two, before the decline resumes. If the evidence has already confirmed the pattern, do not allow a temporary recovery to rewrite the entire story. A bounce back toward the neckline is not automatically proof that the reversal has failed.

While a textbook head-and-shoulders formation appears relatively clean, real market structures can be considerably more complicated. Other minor price formations can develop directly within a larger head-and-shoulders pattern, creating additional localized highs, lows, price gaps, or smaller reversal patterns inside the overarching setup. For example, the right shoulder itself might contain a smaller double top before breaking downward, or a gap down in price might occur during the sell-off from the head. Recognizing these complex configurations becomes easier with experience as you learn to distinguish small internal fluctuations from the overarching reversal structure.

Evaluating the Measured Move

The phrase measured move appears throughout technical analysis and does not always mean exactly the same thing. In general, it refers to an attempt to estimate how far price may travel after some identifiable chart event, such as a breakout or the completion of a pattern. The idea sounds precise, but the reality is much less certain. Measured-move calculations can provide useful reference points, yet historical outcomes vary enough that you should not treat them as guaranteed destinations.

Letting the Pattern Supply the Measurement

One common approach is to use the dimensions of the pattern itself to estimate the size of the move that follows. Consider an ascending triangle, which features a flat horizontal resistance line across the top and a series of rising higher lows along the bottom. To calculate a price target, measure the vertical distance from the pattern's lowest initial low straight up to the flat upper resistance level. If that baseline height measures $5, the traditional rule projects an additional $5 of upside movement applied directly above the horizontal breakout level once price breaks through resistance.

That sounds wonderfully neat. Markets rarely are.

The actual move following a pattern is often smaller than the full height of the formation. Historical studies provide ranges, averages, and modes—the result that occurs most frequently—for different patterns rather than assuming that every breakout travels exactly 100 percent of the pattern's height. Each formation also has its own method of calculating the target. For a head-and-shoulders pattern, for example, you measure the vertical distance between the top of the head and the neckline. You then subtract that distance from the neckline at the point where the downside breakout occurs. The resulting level becomes the estimated target for the decline. Historical studies indicate that this target is reached more than half the time, but that still leaves plenty of occasions when the market stops short or travels much farther.

Resuming the Trend After a Retracement

Another form of measured move occurs when a market makes an initial advance or decline, retraces part of that move, and then resumes the original trend. The retracement may recover or surrender 30 percent of the first move, or some other proportion. What interests the chartist is what happens afterward. The second leg may cover approximately the same distance as the first leg, producing a useful projection.

This concept works especially well within a price channel, where parallel upper and lower boundary lines define an established trend and price swings back and forth between them. To construct a target, identify the first major directional leg within the channel—from its origin point (Point A) to its peak or trough (Point B). Once the price experiences a partial retracement to a new turning point (Point C) along the opposite channel line, measure the precise distance covered during that first leg (A to B) and project that exact amount forward from the retracement turning point (Point C). This calculated extension gives you Point D, your measured objective for the second leg of the trend.

This type of projection has historically reached its target more than half the time in the cited studies. Again, that is a statistical tendency rather than a command from the market. The calculation gives you a place to watch. It does not give you permission to assume that price must get there.

Measuring From a Gap

A third measured-move technique begins with a gap. A gap occurs when the high or low of one bar is separated from the corresponding price range of the preceding bar, leaving a section of prices through which no trades occurred. The gap is significant because something changed abruptly in the market's perception of the security. Buyers or sellers became willing to transact at substantially different prices without filling the intervening area.

The measurement is made by taking the distance from the lowest low of the advance to the middle of the gap and then projecting that distance upward from the midpoint of the gap. In a declining market, the calculation is reversed. The idea is simple enough, but the resulting targets have historically been less dependable than other measured moves.

There is a logical reason for that weakness. A gap is evidence that something unusual has already happened. The traders involved may be reacting to news, fear, uncertainty, changing expectations, or some other event that makes ordinary price behavior less reliable. When the market itself is behaving abnormally, you should be especially cautious about assuming that a neat geometric projection will tell you exactly where it is going next.

A Warning for Today's Charts

You will eventually encounter charts covered with impressive-looking lines and formations that appear to explain everything that happened in the market. Sometimes the underlying analysis is excellent. Sometimes the accompanying commentary is thoughtful and well researched. Yet the actual pattern drawn on the chart may be subjective enough to be almost meaningless.

This is an important distinction. Good analysis does not automatically make a bad pattern good. If an analyst presents an argument you find convincing but draws a formation that does not satisfy the established rules of the pattern, do not simply copy the lines onto your own chart. Rebuild the chart yourself. Identify the highs and lows, apply the pattern criteria, and determine whether the formation actually exists. A pattern should earn your confidence through its price structure and confirmation—not because someone has drawn enough lines around it to make it look convincing.

Drawing Trendlines

Connecting the Dots

People have a strong tendency to look for order in things that are messy. Price charts are no exception. We take a collection of highs and lows that have no obligation to form a straight line and connect them anyway. When we connect a sequence of lows, we create a line that appears to show support, a level where buyers are supposedly waiting to prevent the price from falling farther. Connect a series of highs and we get resistance, the apparent ceiling beyond which buyers have difficulty pushing the price.

This is useful, but it is also dangerous. Price data do not naturally move in straight lines, and the trendline you draw is partly a matter of judgment. There is no guarantee that the line means anything simply because it looks good on a chart. Its usefulness increases when many other traders are watching the same security and independently arrive at approximately the same line. That collective attention can give a trendline practical significance that the line itself does not possess. Because prices generally rise or fall rather than move sideways forever, support and resistance lines are usually angled upward or downward as they follow the trend.

There are other ways to establish these reference points. One of the most popular is the pivot-point method, which takes the high, low, and closing price from a previous period, combines them, and divides the result by three to establish a central level. Additional calculations can then produce related pivot levels. Pivots are often called objective because the calculation comes directly from price data, whereas a hand-drawn support line depends on which lows you decide to connect. But do not confuse mathematical construction with certainty. Pivot levels can be violated by false breakouts just as hand-drawn lines can.

A third approach is linear regression. Rather than selecting a handful of highs or lows, regression attempts to find the line that best represents the entire series of prices over a particular period. As new prices arrive, the line can roll gradually upward or downward to maintain that best-fit relationship. In practical terms, you are dealing with three different kinds of reference lines: hand-drawn support and resistance that tilt with the trend, horizontal levels produced by pivot calculations, and the slowly moving best-fit line created by linear regression. Analysts may draw all sorts of additional lines on their charts, but these three approaches are among the most commonly encountered, and you should understand what each one is trying to accomplish.

Looking Closely at a Price Chart

At first, drawing a trendline seems almost embarrassingly easy. Put your cursor on a low, drag it toward another low, and extend the line into the future. But the real problem is not drawing the line. The problem is deciding whether the line actually represents the behavior of the security. You are looking for a line that accounts for enough of the price action to make it meaningful. If a supposed support line connects only two lows that happen to sit close together, it may be nothing more than a coincidence.

You also need the line to do something useful after you draw it. If price breaks support, you want that break to have enough significance to indicate that the uptrend may be ending. Likewise, when price breaks resistance, you want the move to be strong enough to provide a meaningful indication that the downtrend may be changing. A trendline that looks attractive but produces unreliable signals is not much use to you.

You will find support and resistance lines everywhere in financial newspapers, websites, newsletters, and market commentary. Some are carefully constructed. Others are little more than somebody's opinion drawn with a ruler. You should be skeptical of both until you understand the rules behind the line. A chart covered with trendlines can look authoritative while saying very little.

Following Rules With Rule-Based Trendlines

Rules are useful because they force you to separate what the chart is actually doing from what you want it to do. A rule-based trendline keeps you from conveniently selecting the highs and lows that support your existing opinion. It also gives you greater confidence when entering a rising market or selling into a declining one because you have a defined procedure rather than a vague visual impression.

The rules serve another purpose that is even more important: risk control. A properly constructed trendline gives you a reference point for deciding when the trade is no longer behaving as expected. You do not have to wait until the entire trend has collapsed. The line can tell you when the market has begun to violate the conditions under which you entered.

Drawing Support and Resistance Lines

Support represents latent demand. It is the area where buyers are expected to appear when prices decline. Resistance represents latent supply, where sellers are expected to become more active as prices rise. In an uptrend, support provides the lower boundary of the current price movement; in a downtrend, resistance provides the upper boundary. These definitions are simple, but drawing the lines correctly requires a consistent process.

To construct an uptrend support line, begin with the lowest low in the move and connect it to the next significant low that occurs before a new high is established. As the security continues making higher highs, keep updating the line. Each time a new high is formed, identify the lowest low preceding that high and redraw the support line if necessary. Then extend the resulting line forward into the future using the same slope. Once the security stops producing new highs, stop adjusting the line according to this procedure.

The resistance-line procedure is the reverse. Begin with the highest high and connect it to the next high that occurs before a new low. Continue making adjustments as the market establishes additional lower lows, each time identifying the highest high preceding the latest low. Extend the resulting resistance line forward at the same slope. Once the market stops making new lows, you stop extending the rule-based construction process.

This means trendlines are not permanent objects. They are living parts of the chart. As market conditions change, you may have to erase one line and replace it with another. That can be irritating, particularly when you have grown attached to a line that seemed to work beautifully. The market does not care about your attachment.

Using Support to Enter and Exit

A support line is called support because you expect the market to find buying interest when it reaches the line. Traders who believe the uptrend remains intact may use the test of support as an opportunity to purchase additional shares, and their buying can help prevent the price from falling farther.

Some technicians insist that a valid support line should receive three separate touches before it can be trusted. That requirement has a certain appeal, but it also has an obvious weakness: many genuine trends do not survive long enough to produce a third touch. If you insist on waiting for three touches, you may occasionally watch an excellent trend develop and disappear before you ever enter. Two touches may sometimes be preferable to three because a security that repeatedly returns to the same support level may simply be attracting traders who are determined to test that level again and again.

The basic support-line entry rule is therefore simple: buy on the second or third test of the support line. The exit rule is deliberately less patient. Sell as soon as practical after the low of a price bar falls below the support line. Notice the imbalance. You require evidence before entering, but you react quickly when the trend begins showing signs of failure.

That asymmetry is intentional. You do not want to enter every time price happens to approach support, because some approaches will become breakdowns. Once you own the security, however, your concern changes. You are now protecting capital. A break of support is therefore treated more seriously than an ordinary test of support.

There is an additional complication in some markets, particularly foreign exchange. Traders may deliberately push the price below an obvious support level and then buy aggressively, bringing the closing price back above the line. This can flush out traders who placed stops directly beneath the obvious level. Whether the day's low or the closing price matters more depends partly on the behavior of the crowd trading that particular security. You need to study that behavior rather than blindly applying the same rule to every market.

Repeated tests of support can strengthen your confidence in the line. Each time the day's low reaches the support level without breaking through it, the market demonstrates that buyers are willing to defend that price. This is called a test of support. A successful test can encourage existing holders to maintain their positions and new buyers to enter because they have just seen the support level survive another challenge. Demand increases, and that demand can help propel the next advance.

Noting Breakouts and False Breakouts

A break of support does not necessarily mean that the entire uptrend has ended. You can see a security break beneath its support line and then recover and resume making higher highs. But once the line has been violated, you should no longer treat it as though nothing happened. The line has lost some of its usefulness because the traders watching the security have now seen the violation.

A minor break can therefore mean different things depending on what you are trying to accomplish. From a trend-identification standpoint, a brief penetration of support may not be enough to prove that the trend has reversed. From a risk-management standpoint, however, it deserves your attention immediately. If you had no stop near the support line, the break is a reminder that you need one. If you did have a stop but the price failed to trigger it, reducing part of the position may be worth considering because the original trend is now in question.

A breakout occurs whenever some portion of a price bar penetrates a trendline, although some technicians reserve the word for a closing price that crosses the line. You will encounter the term in many different areas of technical analysis, but the essential idea remains the same: the price has violated a level that traders were using to define the trend. Sometimes the market quickly moves back through the line and continues in the original direction. Even then, the character of the trend has changed because traders have witnessed the violation.

Imagine an uptrend in which the day's low drops beneath support, only for the next several sessions to produce higher highs. The bulls may have effectively rejected the breakdown. What appeared to be a bearish signal was overwhelmed by renewed buying pressure. In another case, the price might spend only one or two days beneath support before returning above it, after which the line resumes functioning as support. This is commonly called a false breakout.

The term false breakout can be misleading. The price really did break the line. What proved false was the assumption that the break represented a lasting change in direction. The distinction matters because you should not become obsessed with deciding whether a breakout is “real” at the instant it happens. What matters is how the market behaves after the violation.

Larry Williams has suggested another way to evaluate a possible false support breakout: examine where the price closed on the day immediately before the violation. In an uptrend, if support is broken but the previous day's close was near the high of that day, the following breakdown may have been caused by temporary profit-taking, a rumor, or ordinary market noise. A close near the low of the previous day presents a different picture and makes a genuine breakdown more plausible. This is not a guarantee, but it gives you another piece of information to consider rather than judging the breakout from a single bar.

Once support has been decisively and repeatedly broken, stop using that line as your primary trading tool. You can leave it on the chart, however. Old support can sometimes become future resistance, and old resistance can later become support. This is one of the oldest observations in technical analysis. It does not happen every time, but the former level can remain useful as a historical reference even after it has ceased to function as the original trendline.

Using Resistance to Enter and Exit

Resistance is the mirror image of support. Draw a line through a sequence of highs and you have a level where buyers have repeatedly failed to push the price higher and where sellers may begin taking profits. Resistance is particularly useful when studying a downtrend because a break above that line can be one of the earliest indications that the declining trend is changing.

There are several reasons to monitor resistance during a downtrend. First, if the decline ends and an advance begins, you want to recognize that change rather than discovering it months later. A break above resistance can be an important clue that the market is preparing to move higher. Second, traders who have the ability to short securities can use resistance as a potential entry area for a short position. This is more accessible in some markets than others; short selling individual stocks can be more restrictive for smaller U.S. investors, while short-side trading is common in markets such as commodities and foreign exchange. Finally, if you already own the security, a break above resistance can validate your decision to continue holding and may provide a reason to add to the position.

Tracking resistance throughout an extended downtrend often requires adjusting your perspective as new price action unfolds. In the initial phase of a sell-off, a downward-sloping resistance line can be drawn across the early rally peaks, offering a distinct short-selling opportunity when price tests that boundary for a third time. If the subsequent drop is aggressive enough, the price may plummet far away from that original trendline, leaving it behind entirely. As the downtrend matures over several months and creates fresh trading waves, the newly established swing highs provide the necessary reference points to construct a second, updated resistance line to track the ongoing decline.

A trader already positioned on the short side might regard the new resistance line as another opportunity to increase the position. A long-only trader sees something different. The second line can serve as a temporary reference for a possible long trade, with the expectation that the price may eventually encounter the older resistance level above it. In the example, selling at the third test of resistance and later covering when the second resistance line breaks would have produced a gain of approximately 20 percent.

The mechanics are the same as support, only reversed. The more frequently the high reaches resistance without crossing it, the more evidence you have that traders are respecting that level. Each successful test tells sellers that buyers are struggling to push through. Sellers may therefore become more willing to act after another failed test.

But resistance can produce false breakouts too. Suppose buyers decide that the security has fallen far enough and is now cheap. Their buying pushes the day's high above resistance, apparently signaling a new uptrend. If other traders do not join them, however, the breakout can quickly lose momentum. The market may retreat back below the line, leaving the trader who bought the initial breakout wondering what happened.

The larger lesson behind support and resistance trading is one of the basic principles of trend following: you are not trying to buy the exact bottom or sell the exact top. Those points are nearly impossible to identify consistently in real time. Your objective is to participate in a meaningful section of the move. Bernard Baruch famously described the idea by saying that he was willing to let other traders have the first and last portions of a move. He wanted the middle. That is a sensible attitude for a trend follower.

Understanding the 1-2-3 Rule

Trendlines require patience because they are not static. You may have to adjust them repeatedly, sometimes from one trading day to the next. The clean textbook example—a rising support line breaks and the market immediately turns into a neat declining trend with a perfectly formed resistance line—is rare. Markets are much messier than that. After breaking support, the price may continue upward for a while, retest the previous high, or move sideways before finally revealing its direction.

To navigate this ambiguity, a structured trend-reversal model breaks the process down into three distinct, sequential price events rather than relying on a single breakout:

1. The Support Break: The initial signal occurs when the price drops below an established upward-sloping trendline. Rather than immediately collapsing into a bear market, the price often continues to fluctuate within a consolidation zone as buyers and sellers fight for control.

2. The High Retest: Following the initial trendline penetration, the market rallies back up to test the previous bullish peak. If buying momentum is truly exhausted, this recovery attempt fails to create a new high, forming a lower peak or a double-top structure.

3. The Low Penetration: As the failed rally rolls over, the price declines and breaks below the lowest reaction point formed immediately after the initial trendline break.

Only when this key intermediate low is penetrated do you have definitive confirmation that the original uptrend has officially ended and a new downtrend is underway.

Sperandeo refers to this as the 1-2-3 method for identifying a trend change. Not every reversal will follow the sequence perfectly, but the concept is valuable because it prevents you from assuming that one trendline violation automatically marks the end of the entire trend. Markets frequently need time to prove themselves.

And if you understand that, you have just encountered one of the most annoying features of technical analysis: the pullback. A pullback is the countertrend or sideways movement that often appears after a breakout. Instead of immediately continuing in the new direction, the price hesitates. Traders move back and forth, positions are adjusted, and the market appears to be stuck.

Congestion is an appropriate description. Imagine a crowded sidewalk in which everyone is trying to move in a different direction. Nobody can make much progress because everyone else is in the way. A market can behave in much the same fashion. Another term for this sideways behavior is consolidation, which describes the period in which market participants appear to be reconsidering and consolidating their views about the security.

Congestion and consolidation are essentially descriptions of a market without a clear trend. Often the boundaries of that range are established by an important previous high or low. Consolidation can appear before a breakout and can also occur immediately after one. If you stare at a short-term chart and cannot determine where the trend is, do not assume the market has stopped making sense. Expand the amount of historical data on the chart and try a longer time frame, such as a weekly chart. The larger trend may become obvious once you step back.

Most technical-analysis techniques ultimately depend upon finding trends to exploit. A market trapped in a narrow, trendless range does not offer the same opportunities to a trend follower. There is nothing wrong with standing aside. In fact, sometimes the correct position is no position at all. A genuine trend follower does not need to be in the market every minute of every day. Wait for the market to provide another trend worth following.

Playing Games With Support and Resistance

Support and resistance are watched by so many traders that they can become partly self-reinforcing. If thousands of market participants identify approximately the same support level, they may all begin buying when price approaches it. Their collective buying then causes the price to respond in the way the chart suggested it would. The line appears to have predicted the bounce, when in reality the behavior of the traders watching the line helped create the bounce.

The same thing can happen on the other side. Some traders may want the price to fall through support so they can purchase the security at a lower price. Larger participants may also have a good idea where less experienced traders tend to place their orders because obvious support and resistance levels are precisely where many small traders put stops and entries. That creates an opportunity for a short-lived move beyond the obvious level.

This is where trading can become psychologically complicated. A professional trader may push price beneath an obvious support level, triggering stops and frightening smaller holders into selling. Once that selling appears, the larger trader can buy at the lower prices created by the panic. If you become too concerned about this possibility, you can end up avoiding every obvious level because you assume someone is waiting to exploit you. Then, in an attempt to outsmart the professionals, you miss the perfectly ordinary entry or exit that the obvious level was providing.

You will also encounter charts where support and resistance appear as beautifully parallel lines. It would be convenient if markets behaved that neatly. Sometimes they do, but often they do not, and even when parallel boundaries appear for a while, they may not remain parallel for long. More importantly, many published charts contain trendlines that were drawn simply to support the author's argument. The line may have little relationship to the rules that supposedly define the pattern.

Do not confuse a chart with a valid analysis simply because it contains straight lines. A good trendline is only the beginning. You also need to study the particular security and the crowd trading it. Does price routinely pierce support by a small amount before recovering? Do traders repeatedly take profits at a certain resistance level? Does a close through the line matter more than the day's extreme? These behaviors differ from market to market.

Your job is not merely to draw a line that looks convincing. Your job is to discover whether the market participants who trade that security actually behave as though the line matters.

Do Prices Pivot?

We know that prices tend to congregate around certain levels for periods of time. The useful question is what happens when that behavior changes. When does the market stop respecting the area where it has been trading and begin moving somewhere else? Pivot points were developed in part to answer that question. A pivot point takes the high, low, and closing price from a previous period, adds them together, and divides the result by three. The period can be a day, a week, a month, or another interval, and the resulting level is then projected forward to provide reference points for the upcoming price action.

From the basic pivot, you can calculate additional support and resistance levels. The first support level is obtained by doubling the pivot and subtracting the previous high. The first resistance level is calculated by doubling the pivot and subtracting the previous low. The second support and resistance levels use the previous trading range as an additional measurement.

The basic formulas are:

Pivot (P) = (H + L + C) / 3

Support 1 = (2 × P) − H

Support 2 = P − (H − L)

Resistance 1 = (2 × P) − L

Resistance 2 = P + (H − L)

To apply these levels in real-time trading, visualize the calculated pivot point as the central anchor, flanked by two lower support levels (S1 and S2) and two higher resistance levels (R1 and R2). As price moves, these stacked horizontal thresholds act as primary decision zones.

For instance, if a declining market drops directly onto Support 1, how you respond depends on your individual risk tolerance. You might treat S1 as a price floor and a logical point to exit a short position, or you might hold out to see if S1 fails, allowing price to extend lower toward Support 2. Conversely, if price bounces upward off support, the overhead resistance pivots (R1 and R2) mark key zones where buying momentum may falter, providing strategic locations to consider new entries or adjust open positions.

Modern traders use pivot points explicitly as potential turning levels. Imagine that the market is falling and has reached S1. You still do not know whether the decline will continue. If price holds S1 and turns upward, that level has provided evidence of support and may offer an opportunity to buy. If price pushes through S1 and continues toward S2, the behavior is different. The market may be establishing a new downward move, and the support level can then become a useful place for risk control.

Some charting programs calculate as many as five pivot levels, and numerous variations exist. The opening price can be incorporated, for example, and different formulas can be used to adapt the levels to particular markets. Thomas DeMark is one of the better-known names associated with pivot analysis. His approach modifies the calculation depending on the relationship between the opening and closing prices. You will find several versions of pivot points in virtually every serious technical-analysis platform.

Pivot points are based entirely on historical data, which gives them an objective quality that hand-drawn trendlines do not have. If two traders use the same look-back period and the same formula, they should arrive at the same levels. A common practice is to calculate the pivots from the previous month and then place those levels on the current daily chart. The calculation is fixed; the interpretation is where the judgment begins.

The method has an interesting history because floor traders used previous-session prices to establish levels they expected to matter during the next trading session. The trader would calculate the levels before the new day began and then monitor the live market in relation to them. Once the pivots were placed on the larger chart, traders could move down to shorter time frames—60 minutes, 120 minutes, and so forth—to see exactly where current price stood relative to the calculated levels.

There is no shortage of educational material devoted to pivot points, including charting tutorials and demonstrations on financial websites and video platforms. The important thing is not to collect twenty different pivot formulas and assume that more calculations will produce better decisions. Learn what the levels represent, observe how the particular market responds to them, and then determine whether they add anything useful to your existing trading method.

Pivot points have one particularly interesting characteristic: they can work well when the market is moving sideways inside a range. In that environment, trend indicators may contradict one another because there is no genuine trend for them to follow. Pivot levels can provide useful reference points when the market repeatedly moves between areas of support and resistance. Their usefulness declines sharply when a major event causes price to move far beyond its normal range. A market driven by an unexpected event is not obligated to respect yesterday's arithmetic.

Drawing Internal Trendlines

Wouldn't it be useful to have a trendline that showed the underlying direction of the market without depending on somebody's decision about which highs or lows should be connected? Hand-drawn support and resistance lines sit along the edges of the price movement. What if you could draw a line through the middle instead?

You can. It is called a linear regression line.

Drawing a straight line through the middle of a series of wildly moving price bars is not something you can do reliably with your eyes. The prices jump around too much. Statistics, however, provides a way to calculate the line that comes closest to the entire series. The calculation finds the line that minimizes the overall distance between the line and the closing prices. This process is called fitting, and the resulting line is the linear regression line. “Linear” refers to the straight-line relationship, while “regression” describes the statistical method used to determine the best fit.

The calculation behind a linear regression can become mathematically involved, but you do not need to perform it by hand. Your charting software will do that work for you. You can study the formula if mathematics is your hobby, but for most traders, learning how the line behaves on an actual chart is considerably more useful. You do not need to understand every component of an automatic transmission to drive a car, and the same principle applies here.

Rules for Drawing a Linear Regression

Unlike a hand-drawn support or resistance line, the regression calculation itself is objective once you establish the beginning and ending points. But you still have to decide where to begin. A logical starting point is a significant recent high or low, particularly the point where the current price movement began.

Of course, that immediately creates another question: What qualifies as recent, and what exactly constitutes the current move? There is no universal answer. For many traders, recent means the previous several weeks or months. I prefer to look for two or more successive higher highs or lower lows, provided the resulting line continues to show a meaningful directional bias. I also generally begin the regression on the bar following an unusually large price bar.

That last rule is partly visual. I consider a bar abnormal when its high-to-low range is roughly 35 to 50 percent larger than the other bars in the current move. The idea is to begin with normal price behavior rather than allowing one exceptional event to dominate the entire calculation. If you are trying to determine the standard behavior of a security, it makes little sense to let a highly unusual bar define that standard.

The amount of historical data you include can make a major difference. Start a regression at a recent low in an orderly price series and you may get a beautiful line, with the closing prices clustered tightly around it. That tight clustering is reassuring. It tells you that the security has been behaving consistently and that future prices may remain reasonably close to the established trend.

Now keep extending the history backward. As more prices enter the calculation, the cluster may begin to spread. The line itself remains mathematically correct, but your confidence in using it as a near-term guide should decline. The farther you reach into the past, the more opportunity you give unusual price behavior to distort the relationship.

Some securities are messy from the beginning. Their price bars may be scattered widely around the regression line, with frequent outliers that appear to have little relationship to the central trend. The regression remains statistically correct—the slope may even be identical to that of a beautifully orderly security—but the dispersion tells you something important. You cannot reasonably expect future prices to stay close to the line. If you are uncomfortable with that uncertainty, use shorter periods, choose securities with more orderly price behavior, or do both.

Where should the regression end? Normally, you stop just before the current bar. You are trying to determine where today's price stands relative to the established trend, so including the current incomplete period can contaminate the measurement. Most charting software will then allow you to extend the regression line into the future. That extension gives you an estimate of where the best-fit trend would place price if the existing relationship continues.

A linear regression does not contain a built-in buy or sell rule. You have to decide how you want to use it. One approach is to trade only securities whose prices remain relatively tightly grouped around the regression line. Another is to reduce or close a position when a previously orderly series begins producing large numbers of scattered outliers. The regression can therefore serve as a measure of both direction and orderliness.

You will sometimes encounter market commentators who claim that because prices have a central tendency, a large departure from the regression line will automatically be corrected by a return to that line. Be careful. That is an attractive story, but it is not a trading system. Prices can remain far away from a regression line for much longer than you expect, and a regression line by itself does not tell you when a return will occur. Treat claims of automatic mean reversion with skepticism.

Identifying Trendedness

Linear regression provides a mathematically precise way to identify and measure the directional slope of price action over a given period. Unlike discretionary trendlines, which require a trader to choose which specific price points to connect, a linear regression line calculates the single best-fit line through a series of price data.

This gives the linear regression an important advantage. For any fixed collection of prices and fixed starting and ending points, there is only one calculated regression line. You do not have to argue about whether one line should connect one low or another low. The calculation determines the answer. Give two traders the same data and the same boundaries, and they get the same line.

That makes a linear regression a useful representation of the underlying trend. If you accept the basic technical-analysis premise that an established trend has some tendency to persist, at least over the short term, you can extend the regression line forward and use it as a forecast of where prices might travel. Some software packages refer to this extension as a time-series forecast.

There are more sophisticated ways to turn historical price relationships into forecasts. Autoregressive integrated moving average, or ARIMA, models are one example. The mathematics is considerably more involved and belongs outside the scope of this discussion. If you want to explore that subject, there are specialized texts devoted to trading systems and statistical forecasting.

For ordinary chart work, however, a simple regression extension can still be useful for testing extraordinary claims. Suppose somebody announces that a $5 stock is going to reach $10 within the next month. Instead of arguing with the prediction, extend the established regression line one month and see whether that destination is even remotely consistent with the existing trend. If the line suggests nothing close to $10 and there is no specific information explaining the expected acceleration, you have reason to question the forecast. A chart can sometimes expose an unrealistic claim without requiring you to predict what will happen yourself.

There is a catch. The regression line depends heavily on where you start and stop the calculation. Change those points and the slope can change substantially. It can point upward, downward, or become nearly horizontal. Applying a single regression line across a complex price structure, such as a V-shaped market bottom, creates significant distortion. While a computer can calculate a single, mathematically valid regression line straight through the center of a V-shaped formation, that single line fails to represent either the initial sharp decline or the subsequent aggressive rally. Because the market contains two distinct directional movements, accurate analysis requires two separate regression lines.

This is why visual fit matters. The less distance between the regression line and the bulk of the price observations, the more useful the line is likely to be as a description of the current trend. A regression can be statistically perfect in the narrow mathematical sense and still be practically useless for trading. The computer has done its job. That does not mean the trader has.

You can therefore create a regression line that is completely valid mathematically and completely worthless as a trading tool. It is still the best fit for the data you gave it, but if the data combine two different trends or contain excessive dispersion, the resulting line tells you little about what prices are likely to do next. Mathematics does not rescue a poor choice of time period.

Using the Linear Regression

Do not treat linear regression as a replacement for support and resistance. It is better viewed as a supplementary tool that can confirm the trend you have already identified. When you choose a sensible starting point and the price remains reasonably orderly around the line, the regression gives you a clean visual representation of the trend. You can see both direction and consistency at the same time.

That second characteristic is easy to overlook. A steeply rising regression line with prices scattered wildly around it tells a different story from a similarly sloped line with prices tightly clustered along it. The first may represent a trend that is unstable or driven by unusual price movements. The second suggests a more orderly market. That distinction can matter when deciding how much confidence to place in a trend-following strategy.

Occasionally, your regression line will fall directly on top of a hand-drawn support or resistance line, or run nearly parallel to it. When that happens, you have an interesting agreement between two different methods of identifying the trend. The hand-drawn line is capturing the edge of the price movement while the regression is capturing its center, yet both are pointing in essentially the same direction. That suggests that the traders participating in the security are behaving in a way that is consistent with the underlying trend.

The opposite situation deserves attention. If your support or resistance line has a dramatically different slope from the regression line, something deserves investigation. One of the lines may be poorly constructed, or the price series may be undergoing a significant change. Likewise, if the regression is climbing at an exceptionally steep angle compared with anything the security has historically produced, the market may be behaving abnormally.

Extreme slopes are rarely comfortable places to make assumptions. A dramatic rise can be sustained for a while, but eventually traders begin taking profits. If the advance becomes sufficiently extreme, profit-taking can turn into panic selling. What looked like strength can then reverse with remarkable speed.

Detecting Abnormal Moves

Linear regression can also help identify price behavior that resembles a bubble. Suppose the regression line begins to steepen sharply, moving far beyond the slope that has characterized the security over the preceding months or years. You can redraw the regression so that it ends at the point where the abnormal acceleration begins. That gives you a representation of the normal trend before the unusual move occurred.

From there, you can extend the older regression forward by hand. The resulting line shows where price might have been if the security had continued behaving according to its previous trend rather than accelerating into the abnormal move. The difference between that projected path and the actual price can help you visualize just how extreme the current movement has become.

There is judgment involved in this process. There is no universal rule saying that a regression slope of a particular angle is officially “too steep.” You have to compare the current behavior with the security's own history and decide when the market has departed materially from its normal pattern.

And remember what you are doing when you extend that line. You are making a forecast, not discovering a fact about the future. The market does not know where your regression line ends. It is simply a statistical representation of what has already happened, extended under the assumption that the previous relationship will continue. Sometimes that assumption is useful. Sometimes the market has other plans.

Transforming Channels into Forecasts

Drawing a single straight line through a collection of prices and then carrying that line into the future creates a problem. Prices do not travel in perfectly straight lines. They wander above and below any trend, sometimes by a little and sometimes by a lot. Rather than pretending that one exact future price can be known, a more useful approach is to establish a probable range. A range is simply the span between the high and low prices over a series of periods, usually days or weeks. Instead of asking where one price will be tomorrow, you are asking where prices are reasonably likely to remain.

There are two principal ways to construct this kind of channel. The first grows naturally out of the support and resistance concepts discussed earlier. The second uses the more statistically oriented linear regression method. Both approaches attempt to accomplish the same basic task: establish boundaries around normal price behavior and then extend those boundaries forward so that you have some idea of what to expect.

Diving Into Channel-Drawing Basics

A channel consists of two straight trendlines that surround a price series. One line runs along the lower portion of the movement and serves as support, while a second line runs along the upper portion and serves as resistance. Ideally, the two lines are parallel. They do not necessarily have to touch every extreme high and low, and the practical rules for drawing them can be less rigid than those used for individual support and resistance lines.

The real purpose of a channel is to organize your perception of the chart. Prices that remain between the two boundaries are behaving as though they belong to the established trend. When prices move outside those boundaries, your attention should immediately increase because the market may be behaving abnormally. The channel therefore gives your eyes a reference. It tells you what ordinary behavior looks like and makes unusual behavior easier to spot.

The Swing Bar Problem

The first question is where to begin. You need a meaningful high or low from which to construct the channel, and that point is usually a turning point in the price movement. Traders commonly call such a point a swing bar.

Unfortunately, the market does not always cooperate by producing a perfectly formed turning point. A security may establish its highest high and then continue drifting upward for several more bars before finally turning down. Likewise, a new low can be followed by several additional attempts at lower prices before the decline actually ends. Eventually, by looking at the placement of the bars, the individual highs and lows, a recognizable pattern, or simply the failure to establish another extreme, you can determine that the move has probably ended.

The word “probably” matters. You can identify a swing point incorrectly, especially when you are trying to do it in real time. Once several bars have passed, the turning point may look obvious. Before those bars exist, it is anything but obvious.

Search the subject online and you will discover a remarkable number of methods for identifying swing bars. One trader wants a certain number of bars on each side. Another adds additional conditions. W.D. Gann used a four-bar approach, and other analysts have built their own variations on the idea. Eventually you can begin to wonder whether there is such a thing as a swing bar at all. There is no single definition that everybody follows.

You are therefore free to begin a channel where you think it makes sense. The important consideration is not whether your swing bar satisfies somebody's favorite formula. What matters is that you choose a starting point that other traders watching the same security are also likely to notice. Remember what you are really trying to do. You are attempting to understand the collective behavior of the market, not win an argument about terminology.

Channels make the chart easier to read because they establish a ceiling and a floor. The resistance line represents the upper boundary where prices may encounter selling pressure. The support line represents the lower boundary where buyers may become interested. Put the two together and you have a projected range for future prices.

Notice the word projected. That distinction is easy to lose when the channel looks unusually clean. A beautiful channel can become psychologically persuasive, and before long you may start believing that prices are required to remain inside it. They are not. A channel is a forecast of probable behavior, not a promise. News, unusual volume, changing sentiment, and random market noise can all send price outside your carefully drawn boundaries.

Drawing Channels by Hand

You may be surprised by how frequently a security will produce a recognizable support and resistance channel. The process is straightforward, although the market will occasionally force you to redraw your work.

Begin by connecting two relative lows toward the lower-left portion of the chart. These lows establish the support line. They should be meaningful relative lows rather than simply two adjacent bars. If another bar sits between them with a higher low, the two points become easier to recognize as separate turning areas.

Once the support line is drawn, extend it into the future using the extension function in your charting program or by dragging the line forward with your cursor. At this stage, the future portion of the line is only a hypothesis. You have not yet seen whether the market will respect it.

Next, wait for a relative high from which you can construct the upper boundary. This is often harder than identifying the original lows. You may think you have found the high, only to watch another bar appear with a still higher high. When that happens, move the resistance line upward. The market gets the final vote.

Extend the resistance line forward once you have established the second boundary. Most charting programs can create a parallel line automatically, which makes the process much easier. Occasionally the parallel resistance line fits the highs beautifully, almost as though the market drew it for you. Do not expect that every time. Naturally occurring perfect channels are considerably less common than charts make them appear in hindsight.

When you first extend these lines, they are hypothetical support and hypothetical resistance. They have not been tested. The distinction is important because traders often begin treating a line as meaningful simply because it looks attractive on the chart.

The line becomes more credible when subsequent price action tests it and fails to penetrate it. If a later high reaches the projected resistance line and retreats, or a later low reaches projected support and turns upward, the market has provided evidence that your extension was useful. A third touch becomes an especially important confirmation point, just as it is with ordinary support and resistance analysis.

Benefitting From Channels

A pair of parallel lines makes a complicated price chart considerably easier to interpret. You have a defined area in which prices are behaving normally and two boundaries that tell you when conditions may be changing. That simple organization can provide several practical benefits.

First, the channel gives you a sense of where prices stand. The upper boundary acts as a ceiling and the lower boundary as a floor. When price reaches one of those boundaries and retreats, you gain additional evidence that the channel is describing the market reasonably well. If the same thing happens repeatedly, you may have a useful framework for anticipating future behavior.

Second, a channel provides a warning when it is violated. A meaningful break tells you that something may have changed in the way market participants perceive the security. You do not automatically know what comes next, but you do know that the old definition of normal behavior deserves another look.

Why Are the Lines Parallel?

The simplest answer is that people like order. Order gives us something to measure against. Latitude and longitude helped sailors establish where they were on an otherwise featureless ocean, and parallel channel lines perform a similar psychological function on a chart.

Still, it is interesting how frequently a support line connecting several lows has a parallel or nearly parallel resistance line that seems to capture many of the important highs. Most charting software therefore includes a command for duplicating or creating a parallel line.

Nobody can point to one universally accepted explanation for why this occurs so often. One possibility is that the crowd itself creates the order. Traders recognize when a security has become relatively expensive near the upper boundary and become reluctant to buy. At the lower boundary, the same security looks relatively cheap, encouraging buyers to step in. Their collective decisions reinforce the channel.

Another explanation is that traders perceive prices as moving cyclically between support and resistance. They buy near the lower boundary and sell near the upper boundary, repeatedly exploiting the range until the pattern changes. That can work, but there is a dangerous temptation to assume the process will continue forever. It will not. Markets eventually escape their ranges.

There is also a less flattering explanation. Analysts sometimes speak about channels as though the lines possess an authority of their own. Then you examine the chart and discover that the lines were drawn rather arbitrarily, touching very few meaningful highs and lows. A channel does not become valid because somebody draws it confidently.

Delving Into the Drawbacks of Channels

The quality of a channel depends heavily on the quality of the price movement it surrounds. In an orderly security, prices may remain close together and the resulting channel can be relatively narrow. That makes deviations easy to recognize.

A disorderly security presents a different problem. When prices jump around repeatedly, the channel must become much wider to contain them. Eventually the boundaries can become so far apart that they tell you almost nothing about what constitutes normal behavior.

There is another practical problem. If you draw an unusually narrow or unusually wide channel that only you can see, there is little reason to expect other market participants to react to it. A forecast range is ultimately a forecast of collective human behavior. The channel has greater significance when other traders are likely to recognize approximately the same boundaries.

That is why the market's reaction is more important than the beauty of the drawing. Your channel can look perfect on the screen and still be irrelevant if nobody else is responding to the same levels.

Channeling to Make Gains and Avoid Losses

If the channel appears to describe the prevailing trend reasonably well, it can provide a straightforward framework for both opportunity and risk control. One approach is to buy near the lower boundary and sell near the upper boundary, repeating the process while the channel remains intact. The strategy works only while the market continues behaving as a range.

The width of the channel also gives you a rough estimate of the amount of money you can reasonably expect to make. Suppose the distance between support and resistance is $5 and you purchase close to support. If the channel remains intact, the next trip toward resistance gives you a probable maximum gain of approximately $5. That is not a guarantee, and you may not be able to enter exactly at support or exit exactly at resistance, but the estimate is useful.

It is useful because it imposes a sanity check. If your trading plan requires the price to travel far beyond the channel in order to produce the profit you expect, something is wrong with the plan. The channel is telling you what the security has recently considered normal.

It also provides a reality check when somebody else makes a forecast. If an analyst predicts a price far above the upper channel boundary, you can compare the prediction with the established behavior of the security. The forecast may eventually prove correct, but the channel tells you that the prediction requires a significant change from the current pattern.

The same logic applies to losses. Once the support boundary is decisively broken, the channel no longer describes the market in the same way. That is the point at which you should have a predefined stop-loss plan. Do not keep defending an invalid channel simply because you liked the trade when you entered it.

Riding the Regression Range

There is another way to construct a channel that is more statistical in nature. Instead of beginning with the highs and lows, you begin with the linear regression line and build boundaries around it. The result is called a linear regression channel.

The linear regression line represents the best-fit relationship among the closing prices. It therefore gives you a central trend, while the channel establishes the range in which prices normally vary around that trend. In that sense, it is a trend channel built from the inside outward rather than from the outside inward.

You use it much like a conventional support and resistance channel. Extend the boundaries into the future and you have a projected range. If prices make a significant break outside one of those boundaries, the old trend deserves to be reconsidered.

The difference in construction is important. A hand-drawn channel begins with the edges of the price movement. The regression channel begins with the center. Once the centerline has been calculated, the upper and lower boundaries are determined from the amount of variation around that line.

Introducing Standard Deviation

Starting with the regression line, you add a certain amount of space above and below it to create the channel. This extra space is based on statistical variation and is commonly expressed through standard deviation. The goal is to encompass a large percentage of the observed price movement while leaving truly unusual highs and lows outside the normal range.

Think of the channel as a fence surrounding the central trend. Most prices should remain inside the fence. An extreme price that travels beyond it represents a larger-than-usual departure from the regression line.

Standard error channels use a related but different statistical calculation. Standard error is concerned with the distance between the observations and their mean, while standard deviation measures variation in another statistical sense. Both approaches can be used to build a channel around a linear regression line, but the padding is calculated differently.

The mathematics behind the distinction is beyond what you need for practical chart reading. Your software may provide one method, the other, or both. Traders frequently use the terms loosely, even though a mathematician would object to treating them as identical. The important trading concept is that both are attempting to establish a statistically probable range around the central trend.

A standard deviation channel is generally narrower than a standard error channel. Gilbert Raff developed a variation of the standard error approach that adds an additional error measurement and therefore produces a wider channel intended to encompass approximately 99 percent of the variations. The advantage is greater coverage. The disadvantage is that the channel can become so wide that it obscures the very trend you are trying to see.

Drawing a Linear Regression Channel

As with the hand-drawn version, the starting point matters enormously. Begin near an obvious high or low and calculate the regression from that point toward a second relative high or low. Your charting software can then construct and extend the channel.

A third test of the channel provides useful confirmation. If another relative high or low reaches the projected boundary and turns away without crossing it, the channel has gained credibility. Sometimes the appropriate turning point is hidden inside an older channel that has already failed. When that happens, return to the bar representing the new turning point and use it as the beginning of the replacement channel.

When constructing channels over a prolonged advance, a security may progress through multiple distinct channels as its volatility or angle of ascent changes. To illustrate this process, consider a stock moving steadily upward through two consecutive channels:

1. The Initial Channel: The calculation begins at the lowest initial low of the advance and extends through the period following the next relative low. At that point, the historical portion of the calculation stops, and the upper and lower channel boundaries are extended into the future to establish a baseline trading envelope.

2. The Extension and Transition: This future extension is hypothetical and serves as an early reference frame rather than a guarantee. As price action continues to evolve, if the security accelerates and breaks out above the original channel's upper boundary, that original channel fails to contain the new trend.

3. Constructing the Replacement Channel: To adapt to the stronger trend, identify the new turning point—the structural low where the price accelerated—and draw a second, steeper regression channel originating from that bar. This secondary channel replaces the old boundary lines and accurately maps the accelerated leg of the rally.

Several months later, the price breaks through the upper boundary. Because the breakout occurs in the same direction as the existing uptrend, the situation deserves special attention. A move like this can represent a genuine acceleration of the trend, but it can also be the beginning of a blowout move that eventually ends badly. The old channel may therefore need to be discarded, although leaving it on the chart can still help you understand how the price behavior changed.

Now construct a new regression and channel from the same starting low, but continue the calculation until another meaningful relative high appears. If the price breaks the upper boundary and then establishes a lower high, that provides evidence that the earlier extreme was in fact a turning point. Extend the new channel forward and watch what happens.

In the example, the market breaks the upper boundary again. Every time price leaves the established channel, uncertainty increases. If you already own the security, that uncertainty may appear favorable because the breakout is upward, but it remains uncertainty. The channel defines what has been normal. Once price escapes it, you have entered a different environment.

What does the second breakout mean? Several possibilities exist. The market may be establishing a new, steeper channel. A blowout breakout may be developing. Or the price may simply retreat and return to the existing channel. The chart alone does not tell you which outcome will occur. This is exactly where another indicator becomes useful. When one tool produces several plausible interpretations, confirmation from an independent measure can help you decide how much weight to place on the breakout.

Confirming Hand-Drawn Channels

You can also use regression channels to test a channel you drew by hand. Begin the regression at the same significant low or high that you used for your support and resistance lines. Then compare the resulting channel with your original drawing.

Sometimes the statistical channel falls almost directly on top of the hand-drawn support and resistance boundaries. That is encouraging because two different methods are identifying essentially the same trend. Your original lines are describing the outer edges of the movement, while the regression channel is using the closing prices to determine the central tendency and its surrounding range.

Do not expect the slopes to be perfectly identical. The hand-drawn channel is based on highs and lows, while the regression calculation relies on closing prices. They are measuring related things, not precisely the same thing.

Sizing Up the Special Features of the Linear Regression Channel

The linear regression channel can be used in much the same way as a hand-drawn channel. You can estimate the probable future range and watch for significant breaks that indicate the established trend may be changing.

There are, however, several characteristics that make it different. First, the regression channel is not designed to contain every price extreme. It is designed to encompass a high percentage of the observations. Some bars will therefore extend beyond the boundaries without necessarily destroying the channel. That is normal statistical behavior.

You can widen the channel by using the Raff method or by instructing the software to include three standard errors instead of the usual two. Doing so can capture approximately 99 percent of the price variation. The trouble is that more coverage also means a wider channel, and an excessively wide channel can make it harder to see meaningful changes in the trend.

Another feature is the ability to make the regression channel self-adjusting. As new data arrive, the software recalculates the channel so that it incorporates the latest information. This may produce a more accurate description of the current data, but it creates a serious problem if your objective is breakout detection. A channel that constantly moves to accommodate the price can make yesterday's breakout disappear from today's chart.

For that reason, a fixed channel is often more useful for forecasting. Draw the original channel, stop it at your chosen ending point, and extend it manually. Then construct a fresh channel from the same starting point through the current period. Compare the two. If their slopes and widths are reasonably similar, your original channel has remained stable. Stability is valuable because it suggests that the relationship you are projecting into the future has not changed dramatically.

A stable channel gives you greater confidence in the forecast embedded in its extended boundaries. It does not make the forecast certain. It simply tells you that the historical structure supporting the forecast has remained reasonably consistent.

Discovering the Drawbacks of Linear Regression Channels

Linear regression and standard error channels have become increasingly popular, but popularity should never be confused with universal usefulness. These tools solve some problems while creating others.

One problem is that there is no guarantee that a large group of analysts will choose the same starting point. The mathematics may be objective once the starting and ending points are established, but the trader still has to decide where those points belong. Put enough analysts on the same chart and you may end up with a collection of competing channels that resembles spaghetti more than technical analysis.

A second problem is that the regression channel does not necessarily stand alone. You can draw one channel over a long period and discover a shorter countertrend inside it. That smaller movement may justify a second channel traveling in the opposite direction. I call these dueling channels because each one can be mathematically correct while offering a completely different interpretation of the immediate market.

This is a signal that another indicator may be needed. A moving average convergence-divergence indicator, or MACD, is one possible tool for resolving the conflict. The point is not that MACD automatically provides the correct answer. The point is that a second independent measurement can help when the channel alone leaves you with competing interpretations.

Consider a market where a security rises inside a broad upward channel over time, yet exhibits multiple brief spikes above its upper boundary and several sharp drops below its lower boundary. Relying solely on the overarching channel boundary suggests the overall uptrend remains intact. However, adding complementary technical indicators paints a conflicting, far more cautious picture:

• Moving Average Crossovers: Within the channel, a short-term moving average crosses below a longer-term moving average immediately following a major upper-boundary spike, signaling a loss of upward momentum.

• Horizontal Line Violations: A separate, hand-drawn horizontal support line across prior swing lows gets tested repeatedly before price finally breaks below it for several trading sessions.

Looking strictly at the raw price action and indicator breakdown—despite the price technically remaining within the broader channel boundaries—you might reasonably conclude that the next major directional movement should be downward.

But another indicator can tell a different story. In the example, the MACD identifies the security as a buy even while price is near the upper portion of the competing channels. Whether you act on that signal is a matter for your trading and risk-management plan. The important lesson is that a channel can organize the evidence without necessarily resolving every disagreement among the indicators.

The third problem is the temptation to call the regression channel “scientific” and then assume that everything about its application must therefore be scientific as well. The calculation is scientific. The software is not guessing about the best-fit line. But you still choose where to start and stop the calculation, and you still decide whether extending the result makes practical sense.

Your car operates according to the laws of physics and combustion, but that does not mean every driver is equally good. The same distinction applies here. The mathematics can be impeccable while the application is poor.

Dealing With Breakouts

The breakout is one of the central ideas in technical analysis because it provides a visible indication that the previous pattern of market behavior may have changed. A price moving outside a channel tells you that the forces controlling the security are no longer producing exactly the same result.

In the simplest interpretation, a breakout means the existing trend has ended in its current form. That does not tell you whether the next trend will be upward, downward, or sideways. It only tells you that you should not expect the security to continue at precisely the same level and rate of movement that existed before the violation.

Breakouts can occur in both hand-drawn support and resistance channels and in statistical regression channels. In either case, a breakout deserves your attention.

But attention is not the same thing as immediate action. You need to determine whether the breakout represents a genuine change or merely a temporary violation of the line.

As discussed earlier, support and resistance levels are watched by many traders. That makes them potential targets for gamesmanship. In an uptrend that is temporarily declining, somebody may push price below support in an attempt to frighten holders into selling. The objective may simply be to acquire the security at a lower price.

The reverse can happen in a downtrend. A burst of buying can force price above resistance and create a new high or a higher close. Short sellers may become frightened and cover their positions, which adds still more buying pressure. The result can look like a meaningful breakout when part of the movement was simply the consequence of traders reacting to the apparent violation.

And sometimes there is no mastermind at all. A breakout can simply be noise, an error in the data, or an unusually random price movement.

Distinguishing False Breakouts From the Real Thing

Suppose your support line sits exactly at $10 and a price bar reaches a low of $9. Has the security genuinely broken support? Maybe. But the answer is not automatically yes.

This is one of those situations where you have to tolerate ambiguity. A channel is an estimate of normal behavior. It is not a physical wall that price cannot penetrate.

Sometimes a price will violate the channel for only a day or two and then return inside the boundaries. Once back inside, the security may resume behaving exactly as it did before. This is commonly called a false breakout.

The terminology can be misleading. The breakout itself was real. Price actually crossed the line. What turned out to be false was the interpretation that the violation meant the entire trend had ended.

In practice, a channel can still accurately describe the overall trading range and prevailing trend even if the price occasionally slips outside its lower or upper boundaries. One or two minor, brief violations should not automatically cause you to discard a well-established channel.

The difficult part is timing. You do not know immediately whether the breakout will become significant or disappear. By the time the answer is obvious, the opportunity to act may already have passed.

The First Line of Defense

Begin with the breakout bar itself. One of the simplest tests is to determine whether the closing price actually crosses the channel boundary or whether only the intraday high or low penetrates it.

The close deserves particular attention because it summarizes the market's final agreement about price for that period. A temporary high or low can be produced by noise or an intraday imbalance. The closing price has survived everything that happened during the bar and therefore often provides a cleaner indication of sentiment.

A support line broken only by the low but recovered before the close tells you something different from a support line that is broken and followed by a close beneath it. The second event represents a more substantial violation.

Verifying With Volume

Volume is another useful piece of evidence. Significant breakouts are often accompanied by noticeable changes in trading activity because a genuine shift in supply and demand tends to attract participation.

An unusually large amount of volume over one or two sessions is called a volume spike. Such spikes frequently occur near the end of powerful advances or declines because buyers and sellers become unusually aggressive. A dramatic rally can end in a burst of buying, while a collapse can produce an equally dramatic surge of selling.

Falling volume can also be informative, although it is more difficult to interpret. Suppose volume has remained relatively steady throughout the trend and then suddenly contracts sharply. Both buyers and sellers may be losing interest. That does not tell you which side will eventually win, but it can indicate that the existing balance is becoming fragile.

Think of it as a logjam. Everyone who wanted to sell at the current price may already have sold, while the remaining holders refuse to sell any lower. The market can sit there until one side finally takes control. When that happens, the eventual move to a new high or low may be accompanied by a noticeable increase in volume.

Looking for Clues From Other Indicators

A breakout is important, but there is no rule saying that one indicator must settle every question. Momentum and relative strength can provide additional evidence about whether the move deserves to be trusted.

For example, a weakening of momentum or relative strength during an uptrend frequently appears before a downside break of the channel. The channel tells you that price has violated its established range. The momentum indicator can help determine whether the underlying force behind the trend was already weakening.

The more independent pieces of evidence that point in the same direction, the easier it becomes to construct a coherent trading decision.

Size Matters, and So Does Duration

You can also use a filter to distinguish a meaningful breakout from a minor penetration. A filter is simply a rule that modifies the raw signal produced by an indicator. Here, the indicator is the channel violation, and the filter determines how far or how long price must move beyond the boundary before you treat the event as significant.

One approach is to require the price to exceed the channel by a specified percentage of the channel's total width. Imagine that the channel is $50 wide. You might require price to move at least 10 percent of that width, or $5, beyond the boundary before declaring the breakout genuine.

Why 10 percent? There is nothing magical about it. Five percent might work better. Twenty percent might work better. The appropriate amount depends on the security and its historical behavior. You can test different thresholds to determine how often the security produces violations of various sizes.

You can also require the closing price, rather than the intraday high or low, to exceed the filtered boundary. This effectively gives the channel some additional breathing room and prevents minor intraday excursions from triggering a signal.

Duration provides another filter. You might allow one bar to violate the channel but require two consecutive violations before declaring the trend broken. Or perhaps you tolerate two violations but not three. You can combine this with the closing-price rule and require the close to move a certain percentage beyond the line for a certain number of periods.

I happen to favor the idea of using something like 20 percent combined with two or three days in some situations, but do not treat those figures as universal rules. They may be completely inappropriate for another security or for another trader's risk tolerance. Each market has its own behavior.

Experienced technical analysts caution against making these filters unnecessarily elaborate. The first reason is that the breakout principle itself is widely recognized. Other traders are watching the same levels, and many of them will respond to a violation without waiting for your carefully engineered collection of conditions.

The second reason is that securities do not all behave alike. A 10 percent filter may have worked repeatedly on one security while another routinely produces 40 percent violations without changing its trend. In earlier decades, a 3 percent filter was commonly used. None of these numbers is universally correct.

The only way to determine whether a particular filter is useful is to examine the historical behavior of the security and test different thresholds. Even then, be careful about becoming overly confident in the result.

There is a third problem: market behavior changes. A security that normally moves in an orderly fashion can suddenly become volatile. A 5 percent filter might work beautifully during a quiet period and then become too small when volatility expands. If you calculate one “ideal” filter from a long historical sample, you may end up with a compromise that is too narrow for volatile periods and too wide for orderly ones.

And there is no way to know in advance which environment tomorrow will bring.

Putting Breakouts Into Context

A genuine breakout means that the old channel has served its purpose. If the violation proves meaningful, discard the channel and begin looking for the new structure. But before deciding that a breakout is genuine, consider the normal volatility of the security itself.

Neatness Counts

A breakout occurring in an orderly price series generally carries different information from one occurring in a highly erratic series. Analysts usually describe this difference with the word volatility, meaning the amount by which prices vary around some central reference. For practical purposes, you can think of low volatility as orderly behavior and high volatility as disorderly behavior.

The more orderly the price movement, the more useful the channel tends to be. A security that normally stays tightly inside its boundaries has established a recognizable pattern, so a decisive escape from those boundaries is more unusual. A security that regularly produces large, erratic bars presents a different problem. Its traders are already accustomed to price wandering well outside the central movement. A breakout that looks dramatic on an orderly security may be routine on a disorderly one.

To understand how volatility impacts breakout signals, compare how a channel behaves across two contrasting market conditions:

• Orderly, Low-Volatility Behavior: Prices fit neatly and tightly inside the upper and lower channel boundaries with minimal random noise. Because the security rarely strays far from its baseline trajectory, any single price bar that breaks cleanly outside the channel stands out immediately as an unmistakable, highly significant event.

• Disorderly, High-Volatility Behavior: Prices jump around wildly, producing large, erratic swings that frequently test or temporarily overshoot the channel lines. If a breakout bar occurs here, it may be the exact same height and dollar size as the bar in the orderly example, but its true significance is much harder to determine because the security already has a routine history of producing wide price excursions.

If a security repeatedly generates false breakouts and those events interfere with your trading, there is a simple solution: find another security. Screening tools can help you locate issues with lower historical volatility and more orderly price behavior.

Transition From Orderly to Disorderly, and Back Again

Changes in orderliness can themselves provide information. When a security shifts from a calm, orderly pattern into a disorderly one, the transition is frequently accompanied by both a channel breakout and a change in volume.

The reverse transition can also be interesting. When a previously erratic security suddenly becomes quiet, volatility falls and buyers and sellers appear to become less willing to act. That temporary inactivity can precede a breakout because the existing balance is becoming unstable.

Then the breakout arrives and participation increases. Volume often expands substantially on the breakout day and during the following one or two sessions. The market has moved from hesitation to action.

Driving Faster Is Always Risky

There is another aspect of a breakout worth considering: where price was sitting inside the channel before it escaped.

Most breakouts occur opposite the direction of the prevailing trend. An upward trend eventually breaks downward, or a downward trend eventually breaks upward. But there are times when price presses against the upper boundary of an uptrend or the lower boundary of a downtrend and then breaks through in the same direction as the existing movement.

That is an acceleration breakout.

Consider a security advancing steadily within a well-defined upward channel. Instead of pulling back toward the central regression line as it normally would after touching upper resistance, the price continues to hug the top boundary. If strong buying momentum forces the price to burst cleanly through that upper channel wall, it signals a sudden increase in the trend's velocity rather than a trend reversal. Do not dismiss this type of breakout merely because price is moving in the "right" direction. It is still a breakout, and it tells you that the behavior of the security has fundamentally changed into a higher-volatility, accelerated state. Sometimes the breakout simply represents a new, steeper trend. The crowd becomes increasingly enthusiastic, buyers chase the security, and the slope accelerates. But an acceleration can also occur near the end of a move. Traders sometimes refer to this as a blowout or blowoff top in an advancing market and a blowout bottom in a declining one.

This creates a counterintuitive situation. An upside breakout during an uptrend can eventually become a warning sign for a downside breakout. The same forces that produce the acceleration can eventually exhaust themselves. The explanation is straightforward. Imagine a security that has been climbing rapidly and attracting enormous attention. Traders become convinced that they must own it before it rises even further. Buying turns aggressive. Eventually, however, everyone who intended to buy for the quick profit has already bought.

Those late buyers are not necessarily long-term investors. Many are simply trying to capture the next burst of upward movement. When the advance finally slows and the chart produces a lower high or lower low, those traders have little reason to remain. They rush toward the exit. When large numbers of traders attempt to sell at approximately the same time, supply suddenly overwhelms demand. The result can be a surprisingly sharp decline. It is much like a market suddenly flooded with tomatoes at the end of a growing season. When there is far more supply than buyers want, the price can collapse.

A declining market can experience the same process in reverse. Eventually, everybody who was determined to sell has already sold. The available supply becomes scarce, while buyers remain interested. To obtain the security, they must bid prices higher until existing holders finally agree to sell.

An upside breakout in an established uptrend can therefore serve as a buy signal, but the opportunity may not last long. The old saying tells traders to buy low and sell high. Larry Williams offered a useful variation: sometimes the trade is to buy high and sell even higher. The important question is whether you recognize when the acceleration is still healthy and when it has begun to turn into exhaustion.

Using Dynamic Lines

Prices do not travel in straight lines, so it makes sense that a useful trend indicator should be able to move with them. That is where the moving average comes in. The moving average is one of the basic workhorses of technical analysis. It is often one of the first tools a new chartist learns, and plenty of traders never move very far beyond it. Look at almost any technical chart and you will probably find one or more moving averages sitting on top of the price bars. Even traders who use completely different methods to select their trades often keep a moving average on the chart because it provides a quick visual reference for the direction and behavior of prices.

A moving average is simply an arithmetic technique for smoothing a series of price observations. One of its advantages is that you do not have to decide where a trend begins or ends, which removes one source of subjectivity that comes with hand-drawn lines. You do, however, have to decide how many periods belong in the average, and that choice matters. We will look at several ways of calculating and applying moving averages, including rules designed to generate buy and sell signals. Keep one point firmly in mind from the beginning: moving averages follow trends; they do not predict them. Any indicator built from a moving average inherits that characteristic. A moving average can continue climbing even after the security has suffered a spectacular decline because the older, higher prices are still buried inside the calculation. The line is telling you what has happened, not what must happen next.

Introducing the Simple Moving Average

The arithmetic is familiar. If you want the average of ten observations, you add the ten numbers and divide by ten. A moving average simply keeps replacing the oldest observation with the newest one. On the next day, the oldest price drops out and today's price enters, leaving ten observations in the calculation. Unless otherwise stated, assume that a moving average uses closing prices. There are exceptions, of course. Ichimoku calculations, for example, use the midpoint rather than the close.

Every serious charting program provides moving averages, and the 10-day version has been around for a very long time. Its popularity dates back to the 1930s, when calculating ten observations by hand was relatively convenient and the result represented approximately two trading weeks. When plotted directly over price action, the value of a 10-day simple moving average becomes clear when compared to a rigid straight trendline. Rather than remaining static, the average bends and dynamically follows the evolving price path while filtering out the random daily spikes that make raw price bars difficult to interpret. It serves primarily as a visual smoothing aid, making the underlying trend direction far easier to identify at a glance.

There is a price to pay for that smoothing. The average contains old information. A 10-day moving average still contains a price from ten trading days ago, so today's line is partly influenced by something that happened nearly two weeks earlier. When the moving average is rising, you know that the newer prices must, in aggregate, be replacing older prices with higher values. The line therefore gives you a picture of trend direction, but it is always carrying some history along behind it.

Starting With the Crossover Rule

When prices trend upward, the moving average normally rises as well. When prices trend downward, the average falls. Because the average is slower than the price, however, the two do not turn at exactly the same time. This creates the simplest moving-average trading rule. Buy when the closing price moves above the moving average and sell when the close moves below it. In actual trading, the transaction would normally be made at the next day's open because you do not know the closing price until the period has finished.

When prices trend upward, the moving average normally rises as well. When prices trend downward, the average falls. Because the average is slower than the price, however, the two do not turn at exactly the same time. This creates the simplest moving-average trading rule:

• Buy Signal: Generated when the closing price moves above the moving average.

• Sell Signal: Generated when the closing price moves below the moving average.

In actual trading execution, transactions are normally executed at the following day's open because the exact closing price is unknown until the trading session ends.

While this mechanical rule captures major directional trends effectively—buying after a market bottom forms and selling after a market top turns down—it carries significant practical risk during non-trending or volatile conditions.

Consider a temporary price spike (an outlier) during an overall trend. The price may close above the moving average, generating a valid buy signal according to the mechanical rule. In real time, a trader has no way of knowing whether this crossover marks the start of a sustained advance or a temporary anomaly. If the price immediately reverses and drops back below the line, the system incurs a whipsaw loss. This illustrates a central vulnerability of moving-average systems: a crossover signal can be completely valid mathematically according to the system rules and still result in a failed trade.

The results in Table 15-1 illustrate the point. The calculation includes short sales as well as long trades. Short selling simply reverses the usual sequence: you sell borrowed shares first and hope to buy them back later at a lower price. Even if you never short stocks yourself, including the short side is useful when testing a trading rule because a sound trend-following method should not mysteriously work only when prices happen to be rising. In the hypothetical example, following every crossover produces a $26.09 gain on $70.61 of starting capital, or approximately 37 percent in less than a year. Buying and holding would have produced only $7.89, or 11 percent. The moving-average method therefore looks impressive, but two of those trades lose money because the apparent crossovers are nothing more than temporary deviations.

Dealing With the Dreaded Whipsaw

A false signal is a signal that looks correct when it occurs but proves wrong afterward. With moving averages, the classic false signal is the whipsaw: price crosses the average, you enter or exit, and then within a few days the price reverses and sends another signal in the opposite direction. Whipsaws are not unusual defects found only in poorly designed moving averages. They are inherent in trend-following. They become especially troublesome when the market moves sideways because there is no sustained trend for the moving average to follow.

Ideally, you want a system that allows you to remain in a profitable position through a large trend while keeping the number of small losing trades under control. That sounds simple until you realize that the two objectives fight each other. A trend-following method makes its money from the large moves. The occasional whipsaw is the price you pay for being positioned early enough to participate in those moves. The trouble begins when the whipsaws stop being occasional. A system that generates one small loss after another can turn into an overtrading machine, producing a great deal of activity without producing a corresponding amount of profit.

Commissions, spreads, and other trading costs make the situation worse. They are not included in the hypothetical calculations here, but in the real world every unnecessary transaction takes something away from your result. A method that constantly jumps from buy to sell and back again can look respectable on a chart while quietly bleeding money in an actual account.

Filtering Out Whipsaws

One solution is to require the crossover to pass an additional test before acting on it. These additional conditions are filters. A filter does not make the moving average less lagging; in fact, it usually makes the signal later. What it attempts to do is prevent you from reacting to every little twitch in the price.

A time filter is straightforward. Instead of buying immediately when price crosses above the moving average, you require the close to remain above the line for a specified number of additional periods. The same idea can be applied to a downside crossover. An extent filter requires price to move a certain distance beyond the average, perhaps a percentage of price or a percentage of the recent trading range. You can make the rule increasingly elaborate by basing the required distance on the recent standard deviation. Whether that extra sophistication actually improves the result is another question.

Volume can provide another test. A crossover accompanied by a substantial increase in trading activity may deserve more attention than one occurring on unusually quiet volume. You can also demand a more extreme price relationship. For example, during an upside crossover, you might require the low of the bar to move above the moving average rather than merely requiring the close to do so. On a downside move, you could require the high to remain below the average. Each additional requirement reduces the number of signals, but every filter introduces another trade-off. You may eliminate a bad signal and, at the same time, eliminate the beginning of a very profitable trend.

Using the Moving Average Level Rule

There is another way to use a moving average that does not require waiting for the price itself to cross the line. Instead, watch the slope of the moving average. An uptrend can be considered finished when today's moving-average value is lower than yesterday's value. Conversely, a downtrend can be considered finished when today's average is higher than yesterday's.

In theory, monitoring the slope can alert you to a slowing trend and get you out of a position earlier than waiting for a full price-crossover event. In practice, however, moving averages inherently accumulate lag because they are calculated from trailing historical data.

Consider a market reaching a sharp peak: after the highest closing price occurs, it can take six trading days for the price to fall far enough to cross beneath a short-term moving average. More critically, it may take as long as ten trading days before the calculated value of the moving average itself turns downward. During that four-day operational delay, price can drop significantly—for instance, suffering an 8.6 percent decline from $82.49 to $75.38 before the level rule formally registers a sell signal. While the overarching trade may remain profitable, this accumulated lag highlights the primary drawback of slope-based trend following.

Accounting for System Performance: Realized vs. Mark-to-Market Gains

When tracking the performance of moving-average systems across historical charts, it is essential to maintain a precise accounting distinction between different types of profit:

• Realized Gain: The actual profit generated from a position that has been closed (sold). The funds are locked in and no longer subject to daily market fluctuations.

• Mark-to-Market Gain: An estimated valuation based on the current closing price applied to an open position. This represents an unrealized paper gain; it fluctuates continuously with price action and can disappear entirely before a formal exit signal triggers.

When evaluating a trading adviser, trading system, or your own results, pay attention to this distinction. Closed-position results tell you what was actually realized. Mark-to-market numbers tell you what an open position is worth at a particular instant. The latter can change immediately.

Dealing With the Limitations

The attractive performance results achieved on a well-behaved chart can easily create the false impression that a moving average is a foolproof trading system. However, thousands of historical price series behave completely differently.

A simple moving average performs best in a strongly trending, orderly market. Aside from an occasional outlier, prices remain reasonably close to the average, allowing a trend-following system to capture sustained directional moves. Put those same moving-average rules into a sideways or range-bound market, however, and the method quickly loses its advantage because a sideways market offers no sustained trend to follow.

When price action becomes highly erratic or volatile, the fundamental trade-off of moving averages is exposed. The indicator is forced to choose between reacting quickly enough to catch a legitimate move and remaining slow enough to filter out meaningless noise:

• Reducing Lag (Faster System): Shortening the look-back period makes the moving average respond quickly to price changes, reducing lag and providing earlier entries and exits. However, this increased sensitivity exposes the trader to constant market noise, causing frequent false signals and costly whipsaws during choppy conditions.

• Reducing Noise (Slower System): Lengthening the look-back period smooths out wild price fluctuations and eliminates minor random spikes. The downside is significant lag, which delays trade signals and causes the system to give back a substantial portion of profits before registering a reversal.

Ultimately, a trader cannot eliminate both lag and noise simultaneously simply by altering the number of periods in a single moving average. Improving execution requires combining the average with complementary filters or adjusting parameters to fit the specific volatility profile of the security.

Fixing Noise

A sideways market cannot be repaired with a moving average because there is no trend for the indicator to identify. Noise is different. If the price series contains frequent abnormal spikes, you can reduce their influence by increasing the number of observations in the average.

Suppose you move from a 10-day average to a 50-day average. A single unusually high or low price now represents a much smaller portion of the calculation. Its influence is diluted by the other 49 observations. That can make the line considerably smoother and less sensitive to temporary price shocks. The disadvantage should be obvious: the signal arrives even later. If ten days of lag already cost you a significant portion of a move, fifty days can be a very expensive cure.

There is another complication. A security can behave neatly for several months and become noisy later, then return to an orderly condition. Changing the moving-average length every time the character of the security changes becomes an endless research project. New data is constantly arriving, so the conditions that justified your chosen period yesterday may no longer exist tomorrow.

One proposed solution is to normalize the price data so that extreme observations have less influence. The normalized value can be expressed as:

Normalized Price = (Current Price − Lowest Price) ÷ (Highest Price − Lowest Price) × 100

If your analysis tells you that a 20-day moving average is appropriate for the current trend, you might also use a 20-day high-low range in the normalization calculation. Some charting programs provide normalization directly, while others allow you to discard or modify unusually large price spikes according to a percentage rule.

But notice what has happened. You have introduced another set of decisions into the moving average. How many periods? What percentage constitutes an outlier? Which look-back period should define the high and low? At some point you may be better off using one of the established moving-average modifications rather than continually altering the raw input.

Fixing Lag

Lag is the other side of the problem. You can watch a security make a powerful move and know perfectly well that the moving average will not respond for several days. By the time your crossover rule gives permission to enter, a substantial portion of the move may already be gone. If you are following the system mechanically, you simply sit there while the potential profit disappears.

The obvious solution is to shorten the moving average. A shorter average responds more quickly because each new price carries greater weight. In general, you want as few periods as possible while keeping the number of false signals within an acceptable range. That sounds like the perfect compromise, but the market charges you for speed.

Applying a three-day moving average demonstrates the extreme end of the sensitivity scale, where the indicator becomes far too responsive to daily price fluctuations. While a three-day average reacts almost instantaneously to direction changes, it abandons the visual smoothing that makes moving averages useful for identifying true underlying trends. Instead of filtering out random market noise, the average mirrors every minor price spike and dip. This hyper-sensitivity generates frequent whipsaws; false crossover signals that trigger premature entries and exits. Applying a three-day moving average to the same price series drastically undermines performance, reducing hypothetical profitability to less than 5 percent. In this instance, hyper-responsiveness performs worse than a simple buy-and-hold strategy, proving that reducing lag at the complete expense of noise reduction is self-defeating.

There is therefore no universally correct moving-average length. The appropriate number depends on the behavior of the security. If the price is noisy enough that you need a long average to suppress the false signals, you must accept the corresponding delay. If that delay makes the strategy unattractive, the answer may not be another mathematical adjustment. You may simply be looking at the wrong security.

Comparing Moving-Average Rules With Donchian Rules

The Donchian entry rule, named after Richard Donchian, one of the early pioneers of technical analysis and managed futures, approaches the problem differently. A common version signals an entry when price reaches a new high over a specified number of periods, often 14 days, and signals an exit when price reaches a new low over the same period. Connecting those successive highs and lows produces something resembling a channel.

One advantage of the Donchian method is that it naturally avoids some range-bound situations that can confuse a moving-average crossover. The price has to establish a new extreme before the system reacts. But the method has weaknesses of its own. A single price spike can create a misleading high or low, and you still have to determine the appropriate number of periods. The argument has simply moved from “Which moving average should I use?” to “How many days should my Donchian look-back contain?”

Magic Moving-Average Numbers

Some traders believe that securities move according to recurring cycles and that the length of a moving average should correspond to one of those cycles. There is some truth behind the idea because market cycles do exist. The problem is that several cycles can operate at the same time, overlap one another, and change in importance. It is rarely possible to identify one fixed cycle and declare that it controls the market.

Belief itself can sometimes become part of the mechanism. If enough traders use the same moving-average period because they believe it corresponds to a cycle, their collective orders can help produce price behavior that appears to validate the cycle. That does not prove that the underlying number is inherently magical.

Commonly watched averages include 14 and 28 days, as well as combinations such as 5, 10, and 20 days. Other traders use combinations such as 4, 18, and 40. The trading month itself contains roughly 20 to 22 sessions, which is quite different from a calendar or lunar month. There is even an old joke among system designers that the four-day average was invented by traders trying to get ahead of those using five days, who were trying to beat the nine-day users, who were attempting to beat the ten-day crowd.

The Ichimoku system provides another historical oddity. Its calculations use 26 days because Japanese markets operated on Saturdays when the method was developed, making 26 sessions a reasonable approximation of a month. Modern Western traders generally leave the number unchanged. That is useful evidence against becoming too attached to the idea that one particular number must represent some universal market rhythm.

The practical lesson is simple: test the data. If a 17-day average happens to fit the behavior of your security better than a 20-day or 14-day average, use the 17-day version. But do not become emotionally attached to it. The market can change, and the period that worked beautifully six months ago may eventually become inferior to 13 days, 23 days, or something else entirely.

One average deserves special attention because of how commonly traders use it: the 20-day moving average. In an orderly, trending security, the 20-day line can behave much like a moving support level. As price approaches it, selling pressure may diminish and buyers may become more active. Less frequently, the same average can act as resistance during a decline. There is nothing mystical about the line. Its value comes partly from its widespread use. Because every trader applying the same calculation to the same closing data gets the same line, the 20-day moving average provides a common reference point that can help you see what other market participants may be watching.

Adjusting the Moving Average

A moving average does not have to be left in its simplest form. You can alter the calculation so that the line reacts more quickly to what prices are doing now while retaining some of the smoothing that makes an average useful in the first place. You will encounter several abbreviations when reading about these indicators. SMA means simple moving average, WMA means weighted moving average, EMA means exponential moving average, and AMA means adaptive moving average. The abbreviations sound more sophisticated than they really are. They are simply different ways of deciding how much importance to give the prices in the calculation.

Folklore Versus Trading Tools

Every so often you will see a headline announcing that a stock or index has moved above its 50-day moving average, slipped below its 100-day average, or that the 50-day average has crossed the 200-day average. The latter is commonly called a golden cross when it occurs to the upside. There is nothing inherently magical about any of these numbers. Their importance comes largely from the fact that many traders watch them. If enough market participants are looking at the same line, the line can become a useful description of the trading environment because traders are reacting to it.

The 200-day moving average is probably the best example of this phenomenon. Even people who otherwise dismiss technical analysis pay attention to it. The 50-day moving average crossing beneath the 200-day average is called a death cross, a dramatic name that sounds far more predictive than the underlying statistic deserves. Historical testing over a very long period has shown that the death cross was followed by lower prices less than half the time. At such long look-back periods, these crossings are better understood as measures of market sentiment and condition than as reliable forecasts. A moving average crossing another moving average is still just a statistical event. The name does not change the arithmetic.

There are many ways to modify a moving average, and the ones discussed here represent only a portion of what has been developed over the years. Do not let the calculations scare you away. You are not expected to sit down with a pencil and calculate these things every evening. Your charting software does the arithmetic. The important part is understanding what the adjustment accomplishes and what you give up in exchange for getting a faster or smoother line.

Weighted and Exponential Moving Averages

One obvious way to make an average more responsive is to give recent prices greater influence. Instead of treating every observation as an equal member of the group, you assign larger weights to the newest observations. Suppose you are calculating a five-day weighted moving average. You could multiply today's price by 5, yesterday's by 4, the price from two days ago by 3, and continue down to a weight of 1 for the oldest price. You then divide by the sum of those weights, which in this case is 15. The result gives today's price considerably more influence than the oldest price.

The exponential moving average, or EMA, takes the same general idea considerably further. The calculation is more complicated, which is one reason you should be grateful that your charting program handles it for you. Rather than simply assigning a fixed weight to each observation, the EMA uses a smoothing factor that pulls the average toward the most recent price. The result is a line that sits closer to current prices than a comparable simple moving average. As the number of periods becomes smaller, the smoothing factor becomes larger and the average responds more aggressively to new information.

Adaptive Moving Averages

An adaptive moving average attempts to solve the problem that has plagued moving-average users from the beginning. You want a short average when prices are moving cleanly in one direction because you want to recognize the trend quickly. But when prices become erratic, you would rather have a longer average that ignores some of the noise and prevents you from being whipped in and out of trades. The difficulty is that you do not know ahead of time whether the next stretch of prices will be orderly or chaotic.

Ideally, the moving average would recognize the change in conditions and adjust itself. You cannot literally change the number of periods from one moment to the next in an ordinary moving-average calculation and expect the resulting series to behave normally, but you can produce a similar effect by changing the way the average responds to price. That is the purpose of an adaptive moving average. When the market is behaving in a manner that calls for sensitivity, the calculation becomes more responsive. When the market is noisy, it places greater emphasis on smoothing.

There are numerous adaptive moving-average variations in modern charting software, many involving some form of detrending. In a simple description, detrending gives the current price greater importance by measuring its relationship to the existing average. Other approaches become much more involved. Perry Kaufman developed one of the better-known versions, the Kaufman Adaptive Moving Average, or KAMA. Other adaptive averages carry names such as AMA, JAMA, associated with Richard Jurik, and HMA, associated with Alan Hull. The names differ, and the mathematics can differ substantially, but the objective remains much the same: reduce lag without turning the average into a noisy mess. Research cited in the source material suggests that adaptive averages do not necessarily provide a dramatic improvement by themselves, but they can still have a useful role when combined with other indicators.

Wild and Woolly Moving Averages

Moving-average experimentation has been going on for decades. In 1960, Chester Keltner described a moving-average system in How to Make Money in Commodities. His method used a 10-day average based on the high, low, and close and constructed a channel around it using the average high-low range. The basic trading idea was straightforward: a move through the upper band generated a buy signal, while a break beneath the lower band generated a sell signal.

Another modification is the triple exponential smoothed average known as TRIX, developed by Jack Hutson. The calculation is considerably more involved than the ordinary moving average, but the practical purpose is easy enough to understand. The smoothing process suppresses many of the minor price changes while giving greater emphasis to larger movements. This reduces some of the lag associated with a conventional moving average and gives TRIX characteristics that make it behave more like a momentum indicator than a plain trend-following average.

As you have probably gathered by now, moving averages are fertile ground for people who enjoy mathematics. There is no shortage of ways to modify the basic calculation. The danger is not in having too few variations. The danger is in believing that another modification automatically produces a better trading result.

Choosing a Moving Average Type

Traders can argue endlessly about which type of moving average is superior. Each calculation method approaches historical price data differently, attempting to balance the trade-off between smoothing out market noise and minimizing lag. Comparing the four primary types reveals how each solves a specific problem while introducing its own set of trade-offs.

The weighted moving average reacts most strongly to the newest prices, followed by the exponential moving average. The KAMA does a particularly good job of filtering the sharp price spikes that make a series noisy, and that can translate into fewer whipsaw trades. But there is a price for that smoothness. A breakaway gap can be treated almost as though it does not exist in the calculation, which means the adaptive average may be slow to recognize a genuine reversal. You can therefore enter a new trend later than you would with a less adaptive method. The reward is that many of the false signals have been removed.

Do not turn a moving average into something it is not. It is arithmetic. It does not possess foresight, intuition, or special knowledge of what the market is about to do. A moving average is a rearrangement and smoothing of the price data. It can help you see the underlying movement, but it is not the movement itself. Important events can occur in the price series that the average barely acknowledges.

Using Multiple Moving Averages

I began with one moving average because a single line makes the basic principle easy to see. In actual trading, however, using only one is uncommon. The short moving average has an obvious attraction: it reacts quickly. The long moving average has an equally obvious attraction: it filters out more of the noise. Why choose between them when you can put both on the chart?

You can use two averages, or you can use three. A short-, intermediate-, and long-term average gives you another layer of confirmation. The following methods show how traders attempt to balance responsiveness against reliability.

Putting Two Moving Averages Into Play

Instead of waiting for the price itself to cross one moving average, you can watch for a short-term moving average to cross a longer-term one. A common example is a five-day average against a 20-day average. The five-day line roughly represents one trading week, while the 20-day line approximates a trading month. Those periods also have practical significance because professional money managers commonly evaluate their performance and positions on weekly and monthly schedules.

The rule is simple. Buy when the shorter moving average crosses above the longer one. Sell when the shorter average crosses below it. Some traders add another condition and exit when the price itself drops beneath the short-term average. There is nothing sacred about 5 and 20, either. You could use 3 and 30, 15 and 24, or almost any other combination. The appropriate parameters depend on the security and the behavior of its prices.

Plotting both the five-day and 20-day moving averages alongside price action demonstrates the key advantage of a dual-moving-average system over a single-average approach:

• Filtering Outlier Price Spikes: A sudden, single-day price spike might force a price-to-moving-average crossover, generating a false or premature signal. With two moving averages, the short-term average absorbs that sudden movement without necessarily crossing above the longer-term average, keeping the system aligned with the broader trend.

• Smoothing Minor Retracements: During an established uptrend, individual closing prices may frequently drop below the short-term average during routine pullbacks. In a dual-average framework, these temporary dips do not trigger premature exit signals because the short-term average itself remains comfortably above the long-term average.

That space between the two averages is useful information. When the averages separate widely, the trend has momentum behind it and the current signal has more room before a crossover can occur. When the averages begin moving toward one another, confidence should diminish. Convergence is telling you that the distance supporting the current trend is disappearing.

Using the two-average method on the example produces a hypothetical gain of $25.31 from an initial stake of $70.61, or about 36 percent. That is below the 61 percent result produced by the moving-average level rule in the earlier example, but the two-average system has several practical advantages. The crossover is visible and can be filtered by waiting an additional day or two or by requiring a specified percentage separation. The system is also less sensitive than the price-versus-average crossover, so it produces fewer false signals at the expense of additional lag. And because it generates fewer trades, commissions and brokerage expenses are reduced.

Trying the Three-Way Approach

If two averages provide confirmation, it is tempting to add another one. A three-average model might use 5, 10, and 20 days. Instead of treating every crossover as a signal, you wait until both the five-day and ten-day averages have crossed the 20-day average before considering the move confirmed. For a trader who buys only and never sells short, another possible rule is to treat a sell signal as occurring when the five-day average crosses either of the other two averages.

This is the belt-and-suspenders approach. You deliberately accept more delay in exchange for greater protection against false signals. Its most useful feature appears when the market stops trending. If prices flatten out or become violently choppy, the three averages may refuse to line up in a way that confirms a trade. Rather than repeatedly entering and exiting as the market chops sideways, you simply remain on the sidelines.

The conventional two-average system does not give you that luxury in the same way. When the averages cross, you exit the existing position and, depending on the rules, may reverse into a short position. In a sideways market, those repeated crossings can turn into a string of whipsaws. The three-average model can avoid much of that damage by requiring an additional layer of agreement before committing capital.

There is a catch. The system is slow. By the time all three averages have arranged themselves into the required pattern, a substantial part of the move may already have taken place. That is why the three-average approach is not nearly as common as the simpler two-average model. You are trading fewer signals for greater confirmation, and sometimes the market has already eaten the best part of the move before you receive permission to participate.

Throw Them All at the Wall and See What Sticks

If two moving averages are useful and three provide more confirmation, why not plot a whole collection of them on a single chart? That is the foundational concept behind the moving-average ribbon. Rather than relying on a pair of isolated indicators, you plot numerous exponential or simple moving averages—often ten or more—across a spectrum of progressive look-back periods (for example, ranging from 4-period to 60-period intervals).

When plotted together, these lines expand and contract dynamically, taking on the visual appearance of a multi-stranded ribbon flowing alongside price action:

• Trend Alignment and Separation: When a market is in a strong, established trend, the individual moving averages fan out and align in parallel order. In an uptrend, the shortest-period averages stack neatly on top, while the longest-period averages trail underneath. The overall width of the ribbon reflects trend strength; a widening ribbon indicates accelerating momentum as short-term and long-term averages diverge.

• Compression and Consolidation: When price enters a sideways consolidation phase or prepares for a major breakout, the ribbon compresses. The individual moving average lines converge tightly into a narrow neck, signifying that short-term and long-term prices are reaching equilibrium.

• Reversal Crossovers: As a trend begins to fail, the shortest-period averages twist and begin crossing through the longer-period averages. Rather than waiting for a single binary crossover signal between two lines, the ribbon provides a continuous visual readout of the reversal as the crossover cascades through each successive moving average line.

By transforming multiple individual trendlines into a unified structural visual, the ribbon allows you to gauge the depth, momentum, and maturity of a trend at a glance. The ribbon becomes particularly interesting when the lines spread apart. The shorter averages are moving faster than the longer ones, creating increasing separation and indicating that the price movement is becoming more strongly organized in one direction. When the lines collapse toward one another, the ribbon is converging and the underlying trend is losing some of its distinctness.

A simple rule is to buy or continue holding only when the current price is above all of the averages in the ribbon. Beyond that, it becomes difficult to construct a clean set of buy and sell rules from the ribbon itself. Its real value is visual. You can see convergence and divergence almost immediately, and sometimes a picture is useful precisely because it does not force you to reduce every observation to one number.

Delving Into Moving Average Convergence and Divergence

A crossover between price and a moving average, or between two moving averages, gives you an event that can be converted directly into a trading rule. But the crossover is rather blunt. If you watch enough charts, you can often see the crossover approaching long before it actually happens. The problem is that a strict rule-based trader must wait for the event rather than act on what appears obvious beforehand.

Look closely at what happens around a turning point. As an uptrend begins to weaken, the short-term moving average starts losing its separation from price. The longer-term average follows later and begins moving toward the shorter average. The two lines are converging. By the time they finally cross, the actual price peak may already be well behind you. The same process occurs at a bottom, only in reverse.

After the crossover, the averages generally move apart. They are diverging. That observation suggests a useful question: what if you measure the amount of convergence and divergence instead of waiting for the actual crossover? Such a measure gives you another way of observing whether the current trend is strengthening or beginning to lose its footing.

The basic principles are straightforward. When two moving averages begin converging, the existing trend may be approaching a change, making convergence an early warning. At a market peak, you can interpret the short-term average losing momentum as evidence that demand is having difficulty producing progressively higher closes even though the longer-term trend remains intact. Near a bottom, a slowing decline in the short-term average can similarly suggest that selling pressure is weakening.

Divergence provides the opposite picture. When there is substantial space between the moving averages, they are moving apart and the current trend is less immediately threatened by a crossover. But even divergence has limits. An unusually large separation can itself become a warning because extreme conditions are rarely sustainable indefinitely. A trend can become so extended that the very strength shown by the wide separation begins to warn of a reversal.

Calculating Convergence and Divergence

The calculation is easier than the terminology makes it sound. Subtract the long-term moving average from the short-term moving average. In an uptrend, the short-term average should normally be above the long-term average, so the result will be positive. Suppose the long-term average is $10 and the short-term average is $15. The difference is $5.

Now suppose prices begin falling. The short-term average weakens first and drops to $13, while the long-term average remains at $15 or is only beginning to respond. Eventually the long-term average also turns down, perhaps reaching $12. The difference between the averages has now narrowed dramatically. What was once a $5 separation has become a much smaller gap. That shrinking difference is convergence.

Gerald Appel built the well-known Moving Average Convergence-Divergence indicator, or MACD, around this principle. His original parameter set used 12-day and 26-day exponential moving averages. You can change those settings in most charting programs, but the 12/26 combination has demonstrated considerable durability over long historical periods. That does not make it magical. It simply means there is a long history of traders using and testing the same basic configuration.

To visualize the indicator, place the 12-day and 26-day exponential moving averages on the price chart itself, and display their calculated difference in an indicator panel directly below. The indicator line is created by subtracting the 26-day average from the 12-day average. When the indicator line rises, the two moving averages are separating (diverging). When it falls, they are moving closer together (converging). When the indicator line touches the zero mark, both moving averages have the exact same value, which corresponds directly to the point where the two moving averages cross on the main price chart.

Creating a Decision Tool

An indicator by itself does not necessarily tell you when to act. To turn the MACD into a more explicit trading tool, Appel added a trigger line: a moving average of the MACD itself, normally a nine-day exponential moving average. The trigger is placed directly over the MACD line, giving you another crossover to watch.

The Moving Average Convergence Divergence (MACD) indicator provides a dynamic, momentum-based alternative to waiting for simple moving averages to cross on a price chart. While standard moving averages are notoriously lagging indicators, the interaction between the MACD line and its signal (trigger) line offers significantly faster responses to shifts in market momentum.

The key advantage of the MACD lies in its ability to generate early actionable signals:

• Early Entry Signals: The MACD line and its trigger line often cross days before the corresponding short-term and long-term moving averages actually cross on the underlying price chart. This head start allows traders to enter emerging trends far earlier than standard trend-following systems allow.

• Trade Preservation on Exit: The lead time provided by the MACD is even more valuable when exiting a position. Because the MACD trigger line turns down well in advance of a primary moving-average crossover, it enables traders to lock in gains and preserve significantly more profit before a major price decline fully registers on lagging trendlines.

• Reentry Opportunities: Following a preliminary exit, the MACD can produce a secondary bullish cross while the broader moving averages remain lagging and uncrossed, providing a clear visual entry signal to rejoin a resuming trend.

Ultimately, the indicator acts as a momentum-driven leading signal for changes already taking place beneath the surface of the price action, giving traders a mechanical edge over traditional lagging averages.

Interpreting the MACD

The MACD histogram offers a visual representation of momentum by translating the distance between two moving averages into vertical bars.

The center line represents the zero mark, where the fast MACD line and the slower signal line cross and equal each other. As market momentum accelerates, the two moving averages pull farther apart; a state known as divergence. On the chart, this widening gap causes the vertical bars to grow taller, visually confirming that the underlying trend is gaining strength.

Conversely, when momentum slows, the moving averages begin contracting back toward one another. The histogram reflects this contraction as the vertical bars progressively shrink toward the zero line. This shrinking pattern provides an early warning that buying or selling pressure is waning, often foreshadowing an impending trend reversal or consolidation before a formal line crossover occurs.

The key advantage of the MACD histogram is readability, as it allows you to gauge changing market momentum at a glance rather than waiting for a delayed crossover signal. However, this flexibility removes mechanical certainty. Because there is no fixed trigger line for the histogram itself, interpreting bar height changes requires discretionary visual judgment.

The zero line is the point at which the two averages are equal. As the histogram bars grow taller, the separation between the averages is increasing, which represents divergence and generally supports continuation of the existing trend. When those bars stop expanding and begin shrinking, the averages are moving toward one another. That contraction deserves attention because it may precede a change in the current signal.

The histogram gives you more freedom, but that freedom comes at a cost. There is no trigger line telling you exactly where to act. You are judging the changing height of the bars with your eye rather than waiting for a specific numerical crossover. This can make the indicator more intuitive, but it also makes the interpretation less mechanical.

The MACD can look almost prophetic when it warns you out of a position well before a major decline, but do not confuse early warning with prediction. The MACD is still constructed from moving averages. It still lags price. A sudden, violent move can outrun it just as it can outrun any other indicator based on historical prices. Nevertheless, the MACD has earned its reputation as one of the more durable and widely used tools in technical analysis.

Gerald Appel's work also illustrates how much staying power a useful indicator can have. Of the numerous books he authored, Understanding MACD, a pamphlet originally published in the 1990s and later reissued with Edward Dobson, became sufficiently difficult to obtain that used copies commanded surprisingly high prices. His broader work, Technical Analysis: Power Tools For The Active Investor, remains another reference for the method.

Modern charting software gives you plenty of ways to modify MACD. You can alter the moving-average calculations themselves, replacing the standard exponential averages with other forms. One variation, DEMACD, substitutes double exponential moving averages in an effort to make the indicator respond more quickly to fresh prices. Another, MACD-V, introduces a volatility adjustment through normalization, with the objective of reducing choppiness and whipsaw losses. The latter received the CMT Association's Charles H. Dow Award in 2022.

You are welcome to experiment with these variations, and your software makes experimentation easy. But remember what you are experimenting with. You are modifying a moving-average-based indicator; you are not transforming it into an oracle. Testing cited in the source material found the basic MACD preferable to DEMACD, but the practical lesson is broader than that particular comparison. A modification should earn its place on your chart through evidence, not because the name sounds more sophisticated or the formula looks more complicated.