Price action

What you remove, and what is left

Nearly every indicator you can put on a chart is computed from price. A moving average is an average of past closes. An oscillator is a ratio between past variations. Those bring in no data the price did not already contain: they transform it, smooth it, rescale it into something readable. That is a useful operation. It is not an extra source of information.

Two measurements escape that rule, and they are worth naming straight away so the claim stays honest: traded volume, and the number of contracts left open. Neither is derived from price, both count something else, and that is why volume can act as a judge where an oscillator can only comment. The market structure lesson uses it to tell a genuine break from a break in an empty room.

One consequence follows, and it is not a matter of opinion. An indicator needs closes that already exist in order to produce its value, so it arrives after the move that produced them. The smoothing that makes it readable is exactly what delays it. The longer you make its period so it stops shaking, the later it speaks to you.

Working with raw price means going back to the source. What is left on screen once you have removed everything comes down to three things: price, time, and the places where price has already done something. The first two read themselves. The third needs sorting, because a year of chart holds hundreds of them and three or four deserve a mark. Everything that follows serves that sorting.

A raw price reading has one property few tools share: it has no settings. No period to choose, so nothing to optimise after the fact on whichever slice of history suits you. That guards you against yourself as much as it saves you time, because a setting always ends up nudged towards what you were hoping to see. What you read this way reads the same on an index, on a currency pair and on a futures contract, on five minutes as on weekly.

A support is not a line, it is a standing appointment

A support is the trace of a place where orders piled up. Price stopped there once because there was something to buy, in a quantity large enough to absorb what was being sold. If some of it remains, price can stop there again. The line you draw holds nothing back: it records where the substance was.

Three different crowds have a reason to act at the same price, and none of them agreed anything with the other two. Those who bought there the first time saw it work and want more of the same. Those who watched the move leave without them are waiting for a return so they do not miss it twice. Those who sold there and got it wrong want out without a loss, which means buying back, at exactly the same place.

Richard Wyckoff, an American operator and publisher of the early twentieth century, gave a name to what builds these congestions: accumulation, and its mirror distribution. His observation still holds. A participant who has to build a large position cannot do it in one go without pushing price against himself, so he works a range, over several sessions. What he leaves behind on the chart is a horizontal area, and that is precisely what you are drawing.

The practical consequence is a harsh filter. Before drawing anything, ask yourself what happened here. If you can answer that price stalled there, that it turned there, that a clean move started from there, the zone has a reason to exist. If the only answer is that price went past, it is not a zone. Most of the lines cluttering a beginner's chart fail that test in two seconds.

Why a band, and never a line to the penny

Nobody is looking at the same price as you. Facing the same low, one person keeps the tip of the wick, another the close of the body, a third the open of the session it happened in. Three honest readings of the same event, a few points apart. Their orders land in three slightly different places, and the sum of those places is a band, not a line.

The plumbing adds to it, and it is craftier than it looks. At any moment there are two prices: the one you can sell at, the lower, and the one you can buy at, the higher. The distance between them is the spread. On most currency platforms your chart draws only one of the two, the lower one. A buy limit sitting on your line is filled only when the other price, the one your chart never shows, comes down to it: the candle has to dip below your mark by the width of the spread before your order goes anywhere.

Add the minimum tick, below which no price exists at all, and slippage on execution as soon as things move fast. Drawing to the penny claims a precision the market gives nobody, and it turns the ordinary width of the order book into something that looks like a contradiction.

What follows changes the way you exit. Price running a few points past your line has not invalidated it: it has swept the handful of orders sitting beyond, which is ordinary market behaviour. A band absorbs that overshoot, a line reads it as an invalidation. You can put a figure on the difference from your own record rather than take my word for it: pull up your last ten losing exits and count the ones where price came back on the right side within two candles.

That leaves the width, which is the real question. It is not counted in fixed points, it is relative to the instrument's usual range on your timeframe. A simple method: take the average range of the recent candles and give your band a fraction of it, enough to contain the wicks that built the zone, no more. ⚠️ A band so wide that everything falls inside it decides nothing any more. If your zone covers a third of the screen, it is not a zone, it is a wish.

The three criteria that qualify a zone

The first criterion is the number of reactions. Intuition says the more the better. Reality is a paradox worth carrying from the start: every visit consumes. Two or three clean reactions mark a zone the market has clearly identified. Eight mark a zone well into its reserves, one that a single further visit may be enough to empty, and nothing on the chart will tell you that this visit is the last one.

The second is the sharpness of the departure, and it is the most informative of the three. Look at how price left the zone the first time. If it left in one go, in one or two wide candles, not everyone who wanted to act had the time to be filled: their intention is intact and it is waiting for a return. If it crawled out over twenty candles, everyone got their fill. The first zone has a reason to hold, the second no longer does.

The third is age, counted in candles rather than in days. Arithmetic says it better than any impression: twelve candles an hour on five minutes, two hundred and eighty-eight a day on a market that does not shut all week, which puts close to fifteen hundred candles between today and last week. The participants who built that zone are long gone. The same stretch of calendar on a weekly chart is a single candle, and the zone is perfectly current. The timeframe decides what old means, never the calendar.

The three criteria do not add up, they filter. A zone that fails one of the three is not half valid, it is doubtful, and a doubtful zone costs more than no zone at all because it works as an alibi when you decide. Write the three answers next to every zone you draw. The ones that pass all three are rare, and their rarity is exactly what you are after.

The spent zone, or why the third touch disappoints

A zone is inventory, not a property of the chart. What stops price is a quantity of resting orders, and that quantity is finite. Every touch spends part of it. Nothing refills it between visits, unless new participants decide to place orders there, which you have no way of knowing in advance.

Three visible signs say a zone is emptying, and none of them needs an indicator. The reaction shrinks: the bounce off the third touch is shorter than the one off the first. The time spent inside grows: price used to take two candles to get out, now it takes seven. The departure slows: the exit candle was wide, it has become narrow. When the three appear together, the zone is at the end of its life.

This contradicts a widespread piece of folklore holding that a level needs three touches to be confirmed. Counting confirms that the market knows the place, which is true and of little use. It says nothing about what remains there, which is the only question at the moment of acting. The third touch often disappoints, for a very precise reason: the first two ate the inventory.

The erasing rule follows from all this, and it is mechanical. A zone crossed with a close on the other side, with no immediate return, has served its time. Erase it. Keeping it because it might work again is the most common way to turn a chart into a grid, and a grid always ends up justifying whatever you already wanted to do.

The polarity flip

A broken resistance often becomes a support, and a broken support often becomes a resistance. This is not a mysterious property of the level nor a courtesy of the market, it is the mechanical consequence of what the break changed for the people who were there.

Unpack it slowly. Those who sold at the resistance are now losing: many will exit at their entry price, which means buying, at exactly the level. Those who watched the break leave without them are waiting for a return to climb aboard, so they buy at the level. Those who bought the break put their protection just under the level, which means they will defend that area before giving up. Three buying flows at the same price, and not one of the three existed before the break.

Robert Edwards and John Magee set this observation down in Technical Analysis of Stock Trends, published in 1948, as an interchange of roles between support and resistance, which usage later settled on calling the principle of polarity. It is one of the oldest regularities the discipline has described, and one of the few whose mechanism can be stated in a paragraph without invoking anything unverifiable. That is probably why it has lasted nearly eighty years.

⚠️ Polarity is in no way automatic, and this is where people get caught. It assumes the break was real: that price spent time on the other side, and that positions were built there. A level pierced by a single wick and immediately reclaimed changed nothing for anybody. No seller is trapped, no buyer missed anything, and the level stays what it was. The test fits in one question: did anyone commit on the other side?

Round numbers, and what humans do with them

Prices ending in round figures concentrate more orders than their immediate neighbours. There is no virtue in the number itself: humans write round numbers. Someone deciding to take a profit writes a price that is easy to say aloud, not a four-decimal figure that fell out of a calculation. Multiply that reflex by thousands of participants and you get clusters of orders in predictable places.

This has been measured, which is what separates the idea from a forum belief. Carol Osler, then an economist at the Federal Reserve Bank of New York, published in 2003 in the Journal of Finance a study of the order books of a large foreign exchange bank. Her result is clean: take-profit orders pile up on round figures, while stop orders pile up just beyond them. Two different behaviours, two different places, and the difference shows in the price.

The mechanism that follows explains a scene you have already watched twenty times. Approaching a round level, price slows: it meets limit orders facing it and absorbing what arrives, a wall that lets itself be nibbled. If it finally gets through, it accelerates: the stops sitting just behind fire and turn into market orders, which do not pick their price and take whatever is there, one level after another. The braking and the surge are not two phenomena, they are the same cluster read from both sides.

⛔ A round level is not a zone in the sense of this lesson. It has no history, nobody built a position there, it qualifies on none of the three criteria. Mark them with a discreet line, use them to understand a hesitation or an acceleration nothing else explains, and never make a round figure the reason for a position. It is a place where something can happen, not a reason to act. Where stops stack around visible highs and lows, and what that changes about placing yours, is the liquidity lesson's job.

The opening gap, and the void it leaves

Sometimes a market reopens a long way from where it closed. The weekend on currencies, the overnight break on shares, the daily halt on futures: wherever a book shuts, there can be a gap. Between the last price quoted before and the first price quoted after, no transaction took place at any price in between. Nobody ever accepted them, because nobody was ever offered them.

⚠️ Do not mistake that void for stillness. Through the break everybody kept revising their price: an earnings release, a number out on Sunday evening, a night of news. The decisions were taken, they simply could not be executed, and the adjustment happens in one block at the reopen. This is the opposite of a sleeping market: it is a market that thought without being able to trade.

The void behaves like the opposite of a zone, which is why it gets handled on its own. In a zone there is substance: orders, positions, committed people. In the gap there is nothing. Nobody holds a reference price there, nobody is trapped there, nobody is waiting for a return. When price comes back into it, there is strictly nothing to slow it down, which explains the speed at which these bands get crossed.

What a gap leaves behind is therefore not the void, it is its two edges. The last price before the break and the first price after were both genuinely traded, both remembered by whoever was there. Those are what you draw, never the middle. The lower edge and the upper edge then qualify like any other zone, on the usual three criteria.

You will often hear that a gap always fills. That is a poor phrasing of a sound observation. A band where nothing was traded is a band with no support underneath, so it is easy to cross, in either direction. This says nothing about when it will happen, nor even that it will: some gaps stay open for years. The difference between the two sentences is the difference between a mechanism and a prophecy.

Three zones you can defend, and tonight's exercise

The familiar trap is drawing fifteen of them. A gridded chart turns any level into a signal: there will always be a line three points from wherever you wanted to enter, and it will agree with you. The number of zones you draw is not a matter of taste, it is what decides whether your chart informs you or acts as your defence lawyer.

The exercise takes twenty minutes and you can do it tonight. Open the daily chart of the instrument you follow. Draw three zones, not four. For each of them, write one single sentence containing three things: why it is there, what would invalidate it, and what you do if price crosses it without slowing down.

A defendable sentence looks like this: zone drawn on a four session congestion, sharp departure and never revisited, invalidated by a daily close under its lower edge, and if price crosses without slowing I erase it and wait for the next reference. A zone you cannot write down that way is not a zone, it is decoration. The test is brutal because it has to be: it removes in three minutes what hours of drawing pile up.

Then comes the upkeep, which is the part almost nobody does. At the start of every session, reread your three sentences and erase the zones price has consumed since. A chart is cleaned like a worktop, otherwise it becomes unusable without you noticing, one line at a time.

Keep those sentences in a file, with their dates. After two weeks you will have around twenty dated zones, some of which held and some of which gave way. Go back over them and look for what the ones that gave way had in common: too many touches, a soft departure, an age outside your timeframe. You will not be taking my word for it, you will have measured it on your own record, and that is the only kind of learning that sticks.