Guide · Chart patterns
Chart patterns in Indian stocks: why they form and how to read any of them
The short answer
A chart pattern is a pause in a trend with structure. Price consolidates because supply and demand have reached a temporary standoff, and the shape of the standoff records who is absorbing, who is exhausted and where trapped positions sit. Patterns fall into three families: continuation, reversal and bilateral. Whatever the name, the tradeable event is the same: a level, a break, a retest, an invalidation and an objective.
Most pattern guides are catalogues: a dozen shapes, a dozen names, a projection formula for each. The names are the least useful part. This page takes the opposite route: first the mechanism, why consolidation happens and what its structure records; then the three families every named pattern belongs to; then the five trade mechanics all of them share. Two anchors keep it honest. The vocabulary comes from the practitioner canon, coined on hand-drawn charts in the 1930s and codified by Edwards and Magee in 1948. The evidence comes from the Journal of Finance in 2000, where the classic shapes were detected by algorithm across three decades of stock data and found to carry some information, and no magic. Both anchors, properly read, point the same way: a pattern is a piece of evidence about supply and demand, and everything that makes it usable sits outside the shape.
Why patterns exist at all
A trend is a persistent imbalance: more aggressive buying than the supply offered to it, or the reverse. No imbalance runs uninterrupted. Winners take profit, opposing conviction arrives at prices it finds attractive, and the aggressor's own demand thins as price moves away from value. When the two sides reach temporary parity, the trend stops and price begins to oscillate. That oscillation is consolidation, and consolidation is the raw material of every chart pattern.
A pattern is a consolidation with structure. Random chop respects nothing; a pattern's swings keep honouring boundaries. A floor that refuses to give way means someone is absorbing everything offered there, test after test. A ceiling that caps every rally means someone is distributing into each push. Converging extremes mean both sides are stepping back. The boundary is the information: it marks the exact price, or the exact trendline, at which one side has repeatedly committed real orders.
Boundaries matter for a second, more mechanical reason: orders concentrate around them, the same logic that builds support and resistance. The defenders of a floor add near it. Traders positioned against the range place stop-losses just beyond it, because that is where their idea is proven wrong. Breakout traders leave entry orders beyond it too, waiting for confirmation. A pattern boundary is therefore not just a line on a chart; it is a shelf of resting orders, stacked and waiting.
That shelf is why breakouts follow through at all. When price crosses the boundary, resting stops convert into market orders in the direction of the break: trapped shorts must buy back, trapped longs must sell out, and waiting entries join them. The follow-through is not enthusiasm; it is obligation. It is also why volume expands on a genuine break, and why the figure below marks the trapped side explicitly: the fuel of a pattern's resolution is other people's exits.
The three families: continuation, reversal, bilateral
Continuation patterns are pauses inside a trend that resolve in the trend's direction. Flags and pennants are the cleanest case: a sharp advance (the pole) creates a crowd of profitable positions; some of them sell; price drifts back against the trend in a narrow channel or a small coil, on fading volume, because the only sellers are profit-takers rather than a new opposing campaign. When that supply is finished, the original imbalance is still in place and the trend resumes. Ascending and descending triangles tell the same story with different geometry: one side keeps stepping toward a level the other side keeps defending, until the defender's inventory runs out. The counter-trend drift inside a continuation pattern is digestion, not a change of opinion. That is the entire meaning of the family.
Reversal patterns record exhaustion. They need a genuine prior trend, because their whole content is that the trending side has stopped being able to extend it; the identical shape inside a sideways range is noise. The head and shoulders is the canonical example: three rallies, the middle one making the final higher high, the third failing below it, with the two intervening lows defining the neckline whose break completes the structure. This hub stays deliberately brief on it, because it has a full ten-diagram guide of its own covering the neckline's slope, volume, the throwback and the inverse version: read the head and shoulders guide for the complete treatment. The double top and double bottom compress the same failure into two touches, a level defended twice and a second attempt that could not improve on the first, covered in depth in its own guide. Rounding tops and bottoms are the slow version: control migrates from one side to the other over weeks, with no single dramatic event, and the only clean signal is the break of the curve's rim.
Bilateral patterns are the family most guides misfile. A symmetrical triangle prints lower highs and higher lows at the same time, which means both sides are retreating from their extremes: neither absorption nor distribution is dominant, and the pattern genuinely does not know which way it will resolve. Calling it bullish or bearish from the shape alone is a guess dressed as analysis. The honest read is narrower and more useful: energy is compressing, resting orders are building on both rails, and the eventual break, whichever way it points, inherits that stored fuel. A bilateral pattern is a promise of movement that stays silent about direction.
| Pattern | Family | What the pause records | The level that matters | Deep guide |
|---|---|---|---|---|
| Flag / pennant | Continuation | A sharp move digesting its own profit-taking | The channel or rail in the trend's direction | Covered on this page |
| Ascending / descending triangle | Continuation, usually | One side stepping toward a level the other keeps defending | The flat boundary | Covered on this page |
| Symmetrical triangle | Bilateral | Both sides retreating; compression without a verdict | Whichever rail breaks first | Covered on this page |
| Head and shoulders | Reversal | Three rallies, the third failing: the trend side exhausted | The neckline | Full guide |
| Double top / double bottom | Reversal | A level defended twice; the second attempt fails | The intervening swing point | Full guide |
| Rounding top / bottom | Reversal | A slow migration of control, no single event | The rim where the curve breaks | Covered on this page |
| Doji, hammer, engulfing | Single-candle events | One session's battle, not a campaign | The bar's own high and low | Candlestick guide |
The canon and the evidence: where patterns come from, and what testing found
The vocabulary is older than most traders assume. Richard W. Schabacker catalogued and named the shapes, head and shoulders and double top among them, across three books published between 1930 and 1934. After his early death, his brother-in-law Robert D. Edwards, with John Magee, codified that work into Technical Analysis of Stock Trends, first published in 1948 and still in print. Every neckline rule and measured-move formula in circulation today descends from that lineage. The pattern names on an intraday Nifty 50 chart were coined on hand-drawn charts of American stocks nearly a century ago, and the fact that they still describe price behaviour is itself worth pausing on: the crowd mechanics underneath them have not changed.
The evidence arrived much later. In August 2000 the Journal of Finance published Foundations of Technical Analysis by Lo, Mamaysky and Wang, still the study serious practitioners cite first. The authors taught an algorithm, kernel regression smoothing, to detect ten classic patterns, head and shoulders, double tops and bottoms, triangles, rectangles and broadening formations, across United States stock data from 1962 to 1996, removing the human tendency to see shapes where none exist. They then compared returns conditioned on a detected pattern against the unconditional baseline. The finding was double-edged and precise: for several of the patterns the conditional distributions differed significantly, meaning the shapes do carry incremental information; and the authors were equally explicit that informativeness is not profitability, which their test did not establish.
Read together, the canon and the evidence agree more than the marketing around patterns suggests. Edwards and Magee's own text is dense with warnings about confirmation, failure and context; the 2000 study found information, not a money machine. A century of practitioner use and the best academic test land on the same sentence: patterns tell you something, and never enough on their own. Any source quoting a precise success percentage for a pattern name is claiming what neither tradition supports. The study ran on American data, but the behaviour it measured, crowds forming and unwinding around recognisable shapes, is not exchange-specific, which is why the honest answer for NSE and BSE charts is the global one: some information, zero certainty, all context.
One trade, five parts: the anatomy every pattern shares
Strip the names away and every pattern trade is assembled from the same five parts: a level, a break, a retest, an invalidation and an objective. This is the section worth memorising, because it transfers: learn it once and you have learned the tradeable half of every pattern in the table above.
The level is the boundary the whole pattern has been arguing about: the flag's rail, the triangle's flat side, the neckline, the twice-defended top. It must be built from real touches. A boundary you had to force through noise is not a level; it is a wish.
The break is a close through the level, not a poke. Indian instruments gap at the open on overnight news, and an intraday spike through a boundary that closes back inside it is the beginning of a trap, not a signal, so completion on NSE and BSE charts is judged on a closing basis. A genuine break should also bring the volume signature with it: quiet inside the pause, expanding at the release, because a real break is resting orders being consumed rather than drift. The breakout guide covers the event itself in detail.
The retest is the market's second opinion. Broken boundaries flip roles: a ceiling that held for weeks becomes the floor the first pullback lands on. A retest that holds converts the break from an event into a structure. A retest that fails is the false break of the next section.
The invalidation is the price at which the pattern's thesis is dead: back inside the range, beyond the retest low, beyond the right shoulder. It must be chosen before entry, because it is the only part of the trade that caps what the market can take from you. The stop-loss is the pattern worn as a number.
The objective is set by the measured move, and its status should be stated exactly: it is a convention, not a law. The convention projects the pattern's height from the break, or the pole for a flag. A range between ₹480 and ₹520 is ₹40 tall, so an upward break carries a reference objective of ₹560. Nothing in market mechanics forces price to honour that arithmetic, and the canon itself offered it as expectation-setting rather than physics. Its real function is discipline: it makes a defined reward exist before the trade does. With a reference invalidation at, say, ₹508 under the retest swing and the objective at ₹560, an entry near ₹520 has a shape, roughly ₹12 of defined risk against ₹40 of intended reward, before a single rupee is committed. Whether that trade should exist at all is a question the pattern cannot answer: the context, the level and the invalidation are upstream judgements, and that upstream work is exactly what the method we teach is built around.
| Check | The question | A pass looks like |
|---|---|---|
| Context | Is there a genuine trend for this pattern to continue or reverse? | A trend visible on the timeframe above the pattern's own |
| Level | Does the boundary connect real touches? | A line you did not have to force through noise |
| Volume | Does participation fade inside the pause and expand on the break? | Quiet coil, loud release |
| Invalidation | What close proves the read wrong? | A price chosen before entry, then honoured |
| Objective | Is the measured move defined before entry? | A reward distance that exists before the trade does |
False breaks: why the obvious level is bait
A false break is a push through the boundary that cannot hold, closing back inside the range. The standard explanation is "failed demand", which is true and incomplete. The fuller mechanism is about liquidity. Everyone reads the same chart, so stops and breakout entries stack just beyond the same obvious levels. That stack is a pool of committed future orders, and committed orders are exactly what a large participant needs in order to transact size without moving the market against itself.
A push through the ceiling makes that pool fire: shorts buy back, momentum entries trigger, and for a moment there is abundant buying at the highs. If the push was genuine demand, the market absorbs the selling and builds on. If it was a seller using the burst as exit liquidity, the buying is consumed, nothing follows, and price falls back through the level. The close back inside the range is the tell, and it is the single most information-dense candle a pattern can print.
The aftermath is why failed patterns travel so well in reverse. The traders who bought the break are trapped overhead, every one of them a future seller, while the shorts who covered into the sweep are gone as buyers. The move that follows a confirmed failure runs on the same stop-cascade fuel as a genuine break, pointed the other way. None of this requires a villain: stop clusters are simply where the liquidity is, and large orders migrate toward liquidity.
The defence is procedural, not predictive. Judge breaks on closes, let the retest speak before adding conviction, and honour the invalidation without negotiation. A trader who cannot stay trapped for long cannot be used as fuel.
The same shapes on every clock
Patterns are fractal. A flag on a five-minute chart and a flag on the weekly are the same object at different magnification, because the behaviour that builds them, absorption, profit-taking, trapped positions, operates at every scale. This is not mystical; it is what happens when the same crowd psychology is sampled at different frequencies.
Reliability is not fractal. Each bar of a weekly chart aggregates the decisions of vastly more participants and capital than a bar of a five-minute chart, so a weekly boundary is expensive to paint and expensive to defend dishonestly, while an intraday boundary can be the work of a single large order. Noise is a larger fraction of small moves, and costs are a larger fraction of small objectives. The same shape means less the faster the clock.
On Indian charts, two practical rules follow. Daily and weekly structures are read on closing prices because instruments gap at the open, and the first and last half hour of the session produce boundary pokes that mean little. A workable hierarchy is to read structure on the daily and weekly charts, and descend to lower timeframes only to time an entry inside a structure the higher timeframe already justified.
The map: where each pattern gets its full guide
This page is the top of the shelf: the mechanism, the families and the shared anatomy. The depth lives in the dedicated guides. The head and shoulders guide walks the full structure in ten diagrams: formation rally by rally, neckline slopes, volume, the measured move, the throwback and the inverse version at bottoms. The double top and double bottom guide gives the twice-defended level the same treatment.
Candlestick patterns are a different shelf, and the distinction matters. A doji, a hammer or an engulfing bar is a single-session event: one bar's battle, resolved by the close. The structures on this page are campaigns built from dozens of sessions. The two layers work together rather than compete: a hammer printed on the retest of a broken ceiling, or a doji at a triangle's final compression, is a one-bar record of exactly the absorption this page describes. Start with how to read candlestick charts, then the single-bar guides such as the bullish engulfing.
If you are building the skill in order: mechanism first, on this page; then the two reversal deep-dives; then candlesticks for reading the individual bars at your levels. Patterns will start looking less like shapes to memorise and more like what they are: the visible footprint of supply and demand exchanging dominance.
Common Questions
Frequently Asked Questions
What are chart patterns?
+Chart patterns are recurring consolidation shapes that price forms when supply and demand reach a temporary standoff inside a trend. The shape records the standoff's structure: a defended floor shows absorption, a capped ceiling shows distribution, converging swings show both sides retreating. Traders classify them into continuation patterns (flags, pennants, most triangles), reversal patterns (head and shoulders, double tops and bottoms, rounding turns) and bilateral patterns (symmetrical triangles). The pattern itself decides nothing; the breakout from it, judged on a closing basis, is the event that carries information.
Why do chart patterns repeat?
+Because the behaviour that builds them repeats. Trends pause when profit-taking meets fresh conviction; participants defend prices where they hold positions; stop-losses cluster just beyond obvious boundaries; and a break converts those resting orders into forced market orders that feed the move. None of that depends on the decade, the exchange or the instrument, which is why shapes named on hand-drawn American charts in the 1930s still appear on Nifty 50 stocks today. Patterns repeat because crowds handle being right, wrong and trapped in the same few ways.
What is the difference between continuation and reversal patterns?
+A continuation pattern is a pause inside a trend that resolves in the trend's direction: flags, pennants and most triangles. The counter-trend drift inside them is profit-taking being absorbed, not a change of opinion. A reversal pattern marks the trending side being exhausted at a level: head and shoulders, double tops and bottoms, rounding turns. It needs a genuine prior trend to reverse; the same shape inside a sideways range is noise. A bilateral pattern, such as the symmetrical triangle, commits to neither side until one rail breaks.
Do chart patterns work on Indian stocks?
+The honest answer is the evidence's answer. The most cited academic test, Lo, Mamaysky and Wang in the Journal of Finance (2000), detected classic patterns algorithmically across decades of US stock data and found that returns conditioned on patterns differed measurably from unconditioned returns: some real information, and the authors stressed that information is not the same as profit. The behaviour being tested, crowds reacting around recognisable shapes, is not exchange-specific, so the same conclusion applies on the NSE and BSE: a pattern is one usable piece of evidence, never a certainty.
What is a measured move?
+The measured move is the convention for setting a pattern's objective: project the pattern's height, or the flagpole for a flag, from the breakout point in the direction of the break. A range between ₹480 and ₹520 that breaks upward carries a reference objective of ₹560. It is a convention for planning, not a law of markets: many moves stop short and some travel far beyond. Its real value is that it forces a defined reward, and therefore a reward-to-risk figure, to exist before the trade does.
What is a false breakout?
+A false breakout is a push through a pattern boundary that fails to hold, with price closing back inside the range. It happens because stop-losses and breakout entries cluster just beyond obvious levels, forming a pool of committed orders; a push into that pool lets large positions transact, and if no genuine demand follows, the level gives way in reverse. The traders who entered the break are trapped, and their exits fuel the opposite move. The defence is procedural: closing-basis confirmation, waiting for the retest, and honouring a pre-set invalidation.
Which chart patterns are most reliable?
+None of them are reliable in isolation, and any source quoting precise success rates for pattern names is claiming what neither the practitioner canon nor the academic evidence supports. Results depend on context: the prior trend, the quality of the level, volume behaviour and the timeframe. What can be said is structural: patterns aligned with the larger trend ask less of the market than reversals; boundaries built from many clean touches carry more information than lines forced through noise; and higher-timeframe structures aggregate more participation. Reliability lives in the checklist, not in the pattern's name.
What timeframe do chart patterns work best on?
+The shapes are fractal: the same structures form on five-minute and weekly charts because absorption and exhaustion happen at every scale. Reliability is not fractal. Each weekly bar aggregates the decisions of far more participants than a five-minute bar, so higher-timeframe boundaries are harder to paint and their breaks mean more. On Indian instruments, which gap at the open on overnight news, daily and weekly breaks are judged on closing prices. A common structure is to read patterns on daily or weekly charts and use lower timeframes only for execution.
Where can I learn each pattern in depth?
+This page is the hub: the mechanism and the shared anatomy. The dedicated guides go deeper on individual structures: the head and shoulders guide covers the full anatomy, neckline, throwback and inverse version with ten diagrams, and the double top and double bottom guide does the same for the twice-defended level. Candlestick patterns, single-session signals such as the doji, hammer and engulfing bar, are a separate layer covered in the candlestick chart guide; they matter most when they appear at the levels and boundaries described here.
Where the facts come from
Sources
- Lo, Mamaysky and Wang (2000). "Foundations of Technical Analysis: Computational Algorithms, Statistical Inference, and Empirical Implementation", The Journal of Finance, Vol. 55, No. 4, pages 1705 to 1765. Establishes the evidence framing used here: kernel-regression detection of ten classic patterns on US stock data from 1962 to 1996 found statistically significant conditional information in several patterns, with the authors explicit that informativeness is not the same as profitability. onlinelibrary.wiley.com
- Edwards and Magee, Technical Analysis of Stock Trends (1948). First edition 1948, building directly on Richard W. Schabacker's books of 1930 to 1934, where the pattern names were coined. Establishes the origin of the vocabulary, the family definitions and the measured-move convention this guide describes.
- Exchange trading mechanics. NSE and BSE instruments can gap at the open on overnight news, and individual stocks trade inside daily price bands. This is why pattern completion on Indian charts is judged on closing prices rather than intraday pokes throughout this guide.
Educational note. This guide explains how chart patterns form and how their structure is commonly read. It is not a recommendation to trade or invest, and it is not investment advice. Bharath Shiksha is an educational publisher, not a SEBI-registered investment adviser or research analyst.