Guide · Price action & market structure

What is price action trading?

The short answer

Price action trading is trading from the price and volume record itself, the open, high, low, close and volume of each bar and the structure they build, rather than from indicators derived from that record. The trader reads swings, levels and the character of moves to judge what buyers and sellers are actually doing, and acts only where a defined level, trigger and invalidation exist.

Most writing about price action starts with pattern names, as if pins and engulfings were the atoms of the market. They are not. Price is the output of an auction: every traded price exists because a market order consumed a resting limit order in an exchange's order book, and every candle is a compressed record of thousands of those events. Read that way, price action stops being pattern astrology and becomes inference: from the shape of the record, you reason about which side is initiating, where resting interest is defending a price, and when that defence has failed. This guide starts at that engine, because the rest of the subject, structure, levels, sequences and the discipline around a trade, only makes sense once you know what a candle actually is.

The engine: what a candle is made of

An exchange holds two kinds of orders. A limit order names a price and rests: buy limits queue below the market, sell limits queue above it, and together they form the order book, a ladder of resting intentions. A market order names no price; it demands immediate execution and is filled by consuming whatever rests on the other side, best price first. The last traded price, the number a chart plots, changes for exactly one reason: market orders on one side consumed all the resting quantity at the best price, and the next fill happened one step deeper in the book. A rally is not price "going up"; it is buyers at market eating through the sell limits at each level faster than sellers replace them.

A candle compresses thousands of those consumption events into five numbers. The high is the deepest level buyers at market managed to push into the offer side during the interval; the low is the same for sellers; the close records which side held the tape when time ran out; volume counts how much resting inventory changed hands to produce all of it. This is why a long upper wick carries information: price reached a level, and the buying that pushed it there met enough resting supply to reverse the tape. Nothing in that reading requires an indicator, because the candle itself is the residue of the auction.

The same mechanics define what traders loosely call support and resistance. A level holds when resting limit orders absorb the market orders thrown at them: volume prints, but price does not step, because the resting quantity refreshes as fast as it is consumed. A level breaks when the incoming market orders eat through everything resting there, and the book steps to the next price. Absorption and consumption are observable events, not metaphors; on a candle chart you see their residue as heavy volume with no progress versus range expansion through the level. The dedicated order flow guide goes one layer deeper, into the depth ladder and the sequencing of individual prints; for price action, the conceptual model is enough.

A level is quantity, not a line. In the left panel market sell orders hit the resting bids at ₹500 and the queue refreshes as fast as it is eaten: volume prints, price does not step, and the candle closes with a wick. In the right panel the same selling consumes the entire queue, the book steps down, and the next trades happen lower. "Support held" and "support broke" are these two outcomes, seen from a distance.

The venue: what NSE fixes about your chart

None of this happens in the abstract. NSE has run a fully electronic market since 1994, when its screen-based NEAT system replaced floor trading, and it matches orders on strict price-time priority: a better-priced order always trades first, and among orders at the same price, the one that arrived earlier does. There is no specialist and no human intermediary who can hold price at a level; every "defence" of a price on an Indian chart is resting limit orders and nothing else. That matters for price action because it means the chart is a complete record of the only mechanism there is. Whatever an institution intends, the only way intention becomes price is through orders in the same book you are reading.

The day's first price is not produced by continuous matching at all. Between 9:00 and 9:08, NSE runs a call auction for the equity market: orders are collected, modified and cancelled without executing, and the entry window closes at a random moment in the final minute to blunt last-second gaming. From 9:08 to 9:12 the system computes a single equilibrium price, the price at which the maximum quantity can trade, with ties broken by the smaller unmatched remainder and then by proximity to the previous close. That equilibrium print is the open on your daily candle. An opening gap is therefore not "price jumping"; it is the auction repricing the stock to clear the overnight order imbalance in one print, and the first minutes of continuous trading from 9:15 are the market testing whether that auction price was right.

The venue also bounds how far a day can travel. Stocks outside the derivatives segment trade inside fixed daily price bands of 2, 5, 10 or 20 percent around the previous close; at the band, orders beyond the limit cannot execute, so the chart flatlines at an administrative price, not a discovered one. Stocks in the derivatives segment carry no fixed band but operate under dynamic price bands that start 10 percent from the previous close and are relaxed in steps during the session, a mechanism SEBI tightened in May 2024 after episodes of disorderly single-day moves. Above all of it, index moves of 10, 15 or 20 percent on Nifty 50 or Sensex, in either direction, halt trading across all equity and equity-derivatives markets nationwide. A price-action reader should know which extremes on a chart are auction outcomes and which are the fence.

Why this section exists. Most price-action articles treat the chart as if it fell from the sky. Knowing that the open is a call-auction print, that a locked upper band is unfilled demand rather than a candle pattern, and that every level is a queue of cancellable orders changes how much weight each mark on the chart deserves.

Market structure: swings and the two regimes

Zoom out from single candles and price organises into swings: a push, a pause, a push. The turning points, swing highs and swing lows, are the units of market structure. An uptrend is a sequence of higher highs and higher lows: each pullback finds buyers at a price above the last one, which is the visible signature of demand stepping up before the old price is even reached. A downtrend is the mirror image, lower highs and lower lows, supply appearing earlier on each rally. The definition sounds trivial. Its power is that it gives "trend" an objective test: an uptrend is intact for exactly as long as the last higher low holds, and not one bar longer.

Structure flips in two acts, not one. First a pullback trades through the most recent higher low, the first time in the sequence that buyers failed to step up early. Then the recovery rally stalls below the prior high, printing the first lower high. Only after both, a broken low and a failed high, has the sequence actually inverted; the zone between them is where the two readings overlap and where positions taken on the old structure sit trapped. Acting on the first event alone means selling every deep pullback in a trend that then resumes. Waiting for both costs some distance and buys a real change in the sequence, which is the trade-off structure reading always presents.

Above the swing level sit the market's two regimes. A range is the auction rotating: price repeatedly rejects both extremes, trade concentrates in the middle, and each probe outside fails for want of follow-through, the market's way of saying both sides accept current value. A trend is the auction migrating: probes in one direction keep finding acceptance, and structure ratchets behind the move. Nearly every price-action technique belongs to one regime or the other. Fading the extremes is a range technique; buying pullbacks to the last breakout is a trend technique. Running one regime's playbook in the other regime is the most common structural error on a retail chart, and the first question a disciplined read settles.

A trend dies in two acts. The uptrend is defined by each low forming above the last. The break of the final higher low is act one; the failure of the recovery rally below the prior high is act two; the shaded zone between them is where longs taken on the old structure sit trapped, and their exits supply the next leg down.

Levels that mean something, and lines that are decoration

A horizontal line earns attention only if something happened there, because the mechanisms that make a level act in the future are positions and intentions created at it in the past. Two dominate. Trapped positions: traders who bought a breakout that failed are underwater from a known price, and a return to that price is their exit; their selling is why the origins of failed breakouts get defended in reverse. Unfilled interest: participants who wanted to buy an area and missed the move leave bids there, or place them when price returns, which is why the origin of a strong impulse so often produces the next bounce. Both mechanisms leave a footprint you can point to on the chart: a rejection, a heavy-volume battle, a gap. A line with no such history behind it obligates nobody, and the market treats it accordingly.

That test separates the levels worth keeping from chart clutter. Prior swing extremes, the previous session's high, low and close, the opening auction print, gap edges and the bases that impulses launched from all mark prices where someone did something and remembers it. A moving average touched twice, a trendline drawn through wicks until it fits, or a retracement percentage of an arbitrarily chosen swing describe past shape without binding anyone in the present. Keep the first list, and make the second earn its place with an actual reaction before you trust it with money.

Levels with memory versus decoration
Level	Why the market remembers it	What invalidates it
Prior swing high or low	The last price where one side overwhelmed the other; stops and breakout orders cluster just beyond it	Acceptance beyond it: consecutive closes past the level with follow-through, not a single wick
Previous day's high, low, close	The reference frame for every overnight position and for the day's gap arithmetic	A session that builds and holds value clearly beyond it
Opening auction print	The equilibrium that cleared the overnight order imbalance; the day's first accepted price	Continuous trade persistently accepting prices away from it
Gap edge	Prices the auction skipped; positions from before the gap sit trapped across it, waiting for a return	The gap filling on expanding range and volume, which releases the trapped inventory
Origin of a strong impulse	Where aggressive orders first hit the book; unfilled interest from those who missed the move rests near it	A return that slices through on expansion instead of producing a reaction
A fitted line with no event	Nothing: it describes past shape; no position was created there, so nobody is obligated at it	Nothing to invalidate: a line that never bound anyone cannot fail

Price action and indicators: the honest relationship

Price action is often sold as a war on indicators. That framing misunderstands what an indicator is. Every standard indicator is a deterministic transformation of the same open-high-low-close-volume series the candles already display. A moving average is price passed through a smoothing filter: it removes detail to expose drift, and pays for it with lag. RSI first differences the closes, then normalises average gain against average loss onto a 0 to 100 scale: it is a statement about the recent sequence of closes, nothing more. MACD subtracts one smoothing of price from another. VWAP is the day's volume-weighted mean price. None of them injects new information; each throws some information away to make one aspect of the series easier to see. That is genuinely useful, and it is also why an indicator can never tell you something price has not already done.

So the practical question is not price action versus indicators; it is whether you can read the source series well enough to know what each transformation is summarising and what it discarded. A trader who understands that RSI "overbought" simply means the recent closes rose fast will not treat it as an automatic fade in a fresh impulse, where fast-rising closes are exactly what initiative buying looks like. Reading price first and keeping a transformation as a summary is coherent. Obeying the summary while ignoring the series it summarises is not.

The source series and its transformations
Series or tool	What it is	What it can tell you	What it cannot
Candles and volume	The auction's raw record: every accepted price and the quantity that traded	Where price travelled, where it was rejected, which side initiated, how much conviction traded	The future, by itself; it records, it does not predict
Moving average	The closes passed through a smoothing filter	The direction and rough slope of drift once noise is averaged out	Turning points as they happen; smoothing guarantees it reports them late
RSI	Closes differenced, then gain normalised against loss on a 0 to 100 scale	How one-sided the recent sequence of closes has been	Whether that one-sidedness is exhaustion or the start of an impulse
MACD	The difference between two smoothings of price	Whether shorter-term drift is gaining or losing on longer-term drift	Anything the two averages do not already contain
VWAP	The session's volume-weighted average price	The average price at which the day's business was actually done	Whether trading above or below it will persist

Reading sequences: impulse and correction

Between the single candle and the full structure map sits the skill that actually separates readers: judging the character of a move while it happens. Directional moves come in two temperaments. An impulse extends range: bars grow, they overlap little, closes land near the extremes, follow-through arrives bar after bar, and volume expands in the direction of travel, all of it the signature of market orders consuming the book faster than it refreshes. A correction is the opposite temperament: ranges contract, bars overlap heavily, closes drift toward the middle, volume tapers, and progress stalls, the signature of a market digesting inventory rather than repricing it. The sequence that defines a healthy trend is impulse, shallow overlapping correction, impulse. When corrections start travelling as far as the impulses, overlap swallows the trend and the reading shifts from trend toward range.

The schematic below walks one sequence. Leg A breaks out on expanding range and holds its gains, five bars with barely any overlap. The correction B overlaps for six bars, gives back well under half of A, and dries up; the shrinking ranges and fading volume mark digestion, not initiative selling. Leg C then extends range again through the prior high, confirming that the demand that produced A is still in force. A trader who can classify A, B and C in real time holds most of what pattern vocabulary is trying to approximate.

Character, not names. Impulse legs A and C extend range with minimal overlap and closes near the highs; correction B overlaps, shrinks and holds far above the origin of A while volume fades. When a "correction" starts travelling as far and as fast as the impulses, it is not a correction any more, and the trend reading is over.

The discipline: a price-action trade is four decisions

Reading is not a trade. A price-action trade exists only when four decisions are on paper before the order goes in. The level: a price with memory, chosen in advance, where you are willing to do business. The trigger: the specific behaviour the tape must show at that level, a rejection, an absorption, a failed probe, that separates your scenario from hope. The invalidation: the price at which the reading is objectively wrong, which is where the stop-loss lives. The size: the risk budget for the trade divided by the distance to that invalidation. Price action without a stop is not analysis; it is opinion.

Numbers make it concrete. A stock near ₹500 breaks a prior swing high at ₹500 on an impulse leg, then pulls back to retest it. The level is ₹500, the old ceiling being auditioned as a floor. The trigger is a bar that probes below to ₹499, finds volume without downside progress, and closes back above ₹502: absorption evidence, not proof. Entry goes at ₹503. The invalidation is ₹496, below the rejection low, because a trade there means buyers did not defend the level and the idea is simply wrong. Risk is ₹7 per share, so a trader whose risk budget is ₹1,750 buys 250 shares and not one more. The first objective is the prior swing high at ₹517, ₹14 away, so the trade risks one unit of capital to pursue two. If the stop is hit, the cost is the ₹1,750 agreed in advance; if the target is reached, the position pays ₹3,500. Both outcomes were priced before entry, which is the entire point.

The bracket is the discipline. The stop sits where the structural reading is wrong, not at a round number of rupees, and the size comes from the stop distance, never the other way around. The same four decisions apply to every price-action setup, whatever the pattern is called.

Where price-action trades break. Three failure modes recur. Entering at the level without a trigger, which is buying a price because it is familiar, not because anyone defended it. Moving the invalidation once price approaches it, which converts a defined risk into an undefined one. And reading a clean structure on one timeframe while the timeframe above is mid-flip, so the "pullback" being bought is the first leg of a new downtrend. The framework catches all three, but only if it is written down before the trade.

Does price action work? The evidence, read honestly

Asked plainly, "does price action work" has no honest one-word answer. There is no public registry of discretionary traders' results, and the venue-level evidence on retail outcomes is sobering: SEBI's FY25 study, published in July 2025, found that 91% of individual equity-derivatives traders lost money, a net ₹1,05,603 crore between them. Those traders were not indicator users as a class or price-action users as a class; they were, overwhelmingly, participants trading without a defined risk framework. That is the correct lesson to take. Pattern names are free, abundant and equally available to everyone, so they cannot themselves be the scarce input. What is scarce is what the rest of this page has described: reading regime before setup, trading only levels with memory, demanding a trigger, and sizing every position from a pre-committed invalidation.

Read that way, price action is best understood not as a strategy but as literacy: the ability to look at the auction's record and say what happened, which side is initiating, and where the current reading fails. Literacy does not produce trades by itself; it produces better-located risk, because a structural invalidation is a fact about the market rather than a number about your comfort. The selection of which setups deserve capital, the process wrapped around execution, and the sizing discipline that keeps any single reading from mattering too much: that upstream work is exactly what the method we teach is built around. The candles are the same for everyone. The framework around them is not.