Loading
Market Intelligence › Basic Terminologies
Structure, trend, levels, candles, volume, indicators — and the limits nobody tells you about
Published by The Retail Trader · JUL 27, 2026 · Price Action
You meet two kinds of people when it comes to technical analysis, and both are wrong. One believes everything is written in the chart, and that if you find the right pattern you can know the future. The other believes it is all astrology and there is no real information in it. The truth sits in the middle, and it is very useful: price carries one thing that you will not find in any balance sheet — what people are doing right now, at this price. That information has its limits, and those limits are very strict. This article holds both things together — what each tool tells you, and where it goes silent.
A price is not an opinion. It is a transaction that has already happened — someone bought at that price and someone sold at that price, and both put their own money behind it. That is why price information is in a different class from any commentary. It tells you what people did, not what they said.
Volume is the second part of that information — how many people did it. And time is the third — when they did it. Put together, these three make a chart, and that chart is really a record of who is stuck where in this name so far, and who is sitting on a profit where.
This is why levels work, as far as they do. A price draws a reaction because many people did something there and their decisions are still alive. When price comes back to that spot, those people act, and that creates a pattern. This is not magic; it is just people's memory.
The most important point first: technical analysis does not tell you what will happen next. It tells you what the situation is right now, and what has usually happened in that situation.
This difference looks small, but it changes the whole game. If you believe the chart predicts the future, every wrong trade feels like a betrayal and every right trade feels like proof. If you believe the chart only describes the situation, a wrong trade is a normal outcome that you had already accounted for.
This is why experienced people always keep an invalidation point alongside a pattern. They know in advance that the setup can be wrong, and they know exactly where it will be proven wrong. People who think in terms of prophecy have no invalidation at all, because in their view the future is already known.
Every setup is a probability, not an outcome. A good setup perhaps works four or five times out of ten, and that is where the money is made — because when it works, it pays more than it costs when it fails.
This means the result of a single trade tells you nothing. A good setup can go wrong and a bad trade can go right. Trying to learn from the outcome is the very thing that keeps most people jumping from one system to another.
So the right way to think about technical analysis is like an insurance company. It does not know what will happen on any one policy. It only knows that across a thousand policies its mathematics works out, and it takes the same risk on each one. A trader's job is the same.
The first decision when you open a chart is the timeframe, and it is the one that gets the least thought. A five-minute chart and a weekly chart of the same name can tell two completely different stories, and both are true.
The choice of timeframe should come from how long you will hold the position, not from your convenience. If you plan to hold a trade for a few weeks, looking for an entry on a five-minute chart takes you into the noise. And if you are an intraday trader, looking for levels on the weekly chart gives you a stop so far away that the position size becomes useless.
A common mistake is switching timeframe when a trade starts going wrong. You took the entry on the five-minute chart, the stop is now getting close, so you open the daily chart and see that it still looks fine there. This is not analysis, it is an excuse. A trade should be judged on the timeframe it was taken on.
Looking at more than one timeframe is useful, but it needs an order, otherwise it just adds confusion.
A working approach uses three timeframes:
• The largest — for direction only. Which way the overall trend is going, and what the big structure is saying.
• The middle — for the setup. This is where you see the level, the pattern and the consolidation.
• The smallest — for entry only. It does not decide the trade, it only improves the timing.
These three have separate jobs and should not do each other's work. The smallest timeframe never decides whether to take a trade or not. If it starts deciding, you will see twenty setups a day and most of them will be rubbish.
And one condition: at least two of the three should point in the same direction. When the larger timeframe's trend is up and you see a downward setup on the smaller one, that trade is possible, but its risk is of a different grade — you are swimming against the current.
A candle is a summary of four numbers — where it opened, where it closed, how high it went, how low it went. The whole journey inside it is not visible, and that is its biggest limit.
The body is the part between the open and the close. The wick is where price went but could not hold. These two things roughly tell you who won during that period.
A long wick is information in itself — price went there and was pushed back. A long lower wick means buyers showed up below. A long upper wick means sellers showed up above. This is an observation, not a signal, and it only says something when it forms at a meaningful level.
A small body says something different — that nobody won during that whole period, and both sides were almost equal. After such periods a direction often emerges, but the candle does not tell you which. People who take it as a direction signal are mistaking an indecision for a decision.
The same data can be shown in many forms, and each form shows something and hides something.
The common forms and their trade-offs:
• Line chart — joins only the closing prices. It looks the cleanest and hides the most, because the whole range of the day disappears. Good for seeing the big structure.
• Bar and candlestick — show four numbers. The most widely used, and enough for most work.
• Smoothed candles such as Heikin Ashi — take averages to reduce noise. The trend looks clear but the price looks wrong, because these are not the real open and close.
A warning is needed on the last one. The price you see on a smoothed chart cannot be traded — it is an average. Taking a stop or entry price from it leads you to a place where the real market does not exist. It is fine for looking, not for orders.
And constantly changing the chart type creates the same problem that constantly changing the timeframe creates. A trade taken in one form should be judged in that same form.
On a long-term chart, a linear scale misleads. On it, the move from one hundred to two hundred and the move from two hundred to three hundred look like the same distance, even though the first is a doubling and the second is only fifty percent.
So a chart spanning several years should be viewed on a log scale. On it, equal percentage changes look like equal distances, and only then do old trendlines mean anything. On an intraday chart or one covering a few weeks, this difference is almost nil.
The simplest and most useful technical idea is structure, and it needs no indicator. Price is either making a steady series of higher highs and higher lows, or a steady series of lower ones, or nothing in particular.
The first condition is called an uptrend. Each fall stops above the previous fall, and each rally goes beyond the previous rally. It means buyers are arriving a little earlier each time and sellers a little later.
The second condition is the opposite. And the third, where there is no clear pattern, is the most common — the market spends most of its time in it, and this is where the most losses happen, because people keep applying trend tools in it.
An uptrend breaks when, for the first time, a fall goes below the previous fall. This is a clear, objective thing — there is no opinion in it, just a comparison of two numbers.
For this one reason, structure is the most dependable tool. It leaves you no room to ask whether the trend is still there; it gives you one price, and when price goes below it, the answer changes.
But there is a condition — structure always exists on one timeframe. A trend can break on the daily while it is still intact on the weekly. This is not a contradiction. It only means the shorter-timeframe players are leaving while the bigger ones are still sitting.
The whole matter of structure rests on swing points — the highs and lows that make up the trend. And everyone has their own way of counting them, which is how two people looking at the same chart can see different structures.
The way to avoid this is a condition fixed in advance. A common condition is that a swing high is a point with a few lower candles on both sides of it. How many candles is your decision, but you make that decision once and keep it the same on every chart.
The benefit is that structure turns from an opinion into a rule. You do not ask what you feel; you check whether the condition was met. On bad days, this is the one thing that keeps decisions steady.
And one small point that makes a big difference — stop counting very small swings. If you treat every two-candle jerk as a swing, structure will appear to break every day and that information will be worthless. The whole value of structure lies in the fact that it changes only occasionally.
When price keeps moving between two levels, it is called a range. A range is more common than a trend, and the way of trading it is exactly the opposite — in a trend you buy the breakout, in a range you get trapped on the breakout.
The simple logic of working inside a range is that there are sellers at the top and buyers at the bottom, and as long as they are there, price will keep moving around in the same place. A range ends when one of those sides runs out.
The costliest mistake is to mistake a range for a trend. Inside a range, every small rally looks like a breakout and every fall looks like a collapse. People who react to every one of these get cut twenty times inside the range, and this is the most common form of loss in retail accounts.
A trend does not always run the same way. Roughly three phases can be seen in it, and the experience of trading each one is completely different.
The three phases of a trend:
• The start — when the structure has only just changed. This has the best risk-reward and the fewest people believe there is a trend.
• The middle — when it has begun to look obvious. Following the trend is easiest here and most of the money is made here.
• The end — when it has become visible to everyone. Moves are fast here, volatility rises, and risk-reward is at its worst.
Recognising each phase is easy afterwards and hard at the time. One useful hint is whether talk of the trend has become commonplace — when it is visible everywhere, the last phase is often under way. This is not a timing tool, just context.
Taking a trade against the trend appeals to everyone because of the thrill of catching a top or bottom. Its mathematics is harsh and almost always ignored.
In a trade with the trend, you stand in the direction in which more people are arriving. Against it, you have to be right and also get the timing right, because a running trend can take out your stop before you are proven right.
If you do want to take a reversal trade, one condition can be applied to it — let structure break first. This is a wait that does not give you the very top of the move, but it saves you from the ten times you would have put your hand under a falling knife.
A level forms because something happened there before. Either a lot of trading took place there, or a sharp move started from there, or price stopped there again and again.
The psychology behind it is simple. Those who bought there and then watched price fall want to get out when it comes back — so when price arrives from above, that level becomes resistance. Those who missed their chance to sell there want another chance. And those who bought there want to buy again.
This is why old levels are remembered for so long. They were not formed by any formula, they were formed by people's trapped decisions, and those decisions stay alive until they are closed.
It is easy to draw a thin line on the chart for a level, and that is the biggest misunderstanding. A level is always an area — a band a few rupees wide above and below.
The reason is that the level was not formed by any one trade. It was formed by many trades made over a period, and those trades took place over a range. So expecting a reaction at one exact number is wrong.
The practical effect is on the stop. Placing the stop right next to the level means getting cut inside the zone itself. The stop should be beyond the zone — where it is fair to say the level broke, not merely that the level was touched.
When a support breaks, it often becomes resistance, and vice versa. This is the best-known technical idea and its logic lies in people's positions.
The people who were buying at that support are now trapped. When price comes back to that level, many of them get their money back and leave — and that selling is what stops the level from going up now.
This does not happen every time. If very little trading took place at that level, or it is very old, there are very few trapped people there and the reaction will be weak. So the strength of a level comes not from its age but from the volume that traded there.
One more thing that weakens this effect is arriving from very far. If price reaches that level after a long fast move, many people have already left on the way, and the pressure remaining at that level is not as great as it looks.
A trendline is a slanted line joining a series of rising lows. Its logic is the same as a horizontal level, only it moves along with price.
Its biggest weakness is that it is very subjective. Three different trendlines can be drawn on one chart and all three will look right. That is why trading off a trendline is weaker than trading off a horizontal level — too much of your own preference goes into it.
If you do use a trendline, a few conditions make it better:
• It is built from at least three points — any line can be drawn through two points.
• Whether you touch the wicks or the bodies, whichever you choose, keep it the same across the whole chart.
• Trust a very steep line less — it almost always breaks quickly.
• The break of a line is not a signal in itself, unless structure breaks as well.
A channel is a pair of parallel lines — one above the trend and one below. Its most useful purpose is to set realistic target expectations, not for entry. If price is moving inside the channel, the distance to the other edge is a reasonable expectation; expecting much more than that often turns out wrong.
Big round numbers form a faint level of their own, because people place orders there. There is no deep reason — it is just a habit, but it is the habit of many people, so its effect shows up.
The most useful application is the reverse. If you need to place an exit or entry right near that round number, placing it a little before often gets a better fill, because there is a crowd there. It is a small thing that adds up to something heavy for people who trade often.
Behind the name of every candlestick pattern there is one thing — which side's people got tired and which side's people arrived. Understanding this is more useful than memorising names.
A candle with a long lower wick is saying that selling happened below but buying came in that was bigger. A big red candle that swallows the whole previous green candle is saying that more selling happened in this one period than all the buying in the previous period.
In the same way, every pattern can be read through the transactions inside it rather than through its name. People who read it this way do not need to memorise pattern names at all, and they can read new patterns on their own.
This is the most important line of the whole section. The same candle says a lot in one place and nothing at all in another.
A reversal candle after a long trend, at a meaningful level, on higher volume — that is information. The same candle in the middle of a range, with no level, on ordinary volume — that is just noise.
Before trusting a pattern, ask these three questions:
• Where did this pattern form — at a meaningful level, or somewhere in the middle.
• What happened before it — a reversal pattern needs a trend first, otherwise what is it reversing.
• What volume said — whether participation rose with the pattern or not.
Most people do only the first job — recognising the pattern. The other two questions are what make a pattern worth trading, and asking them requires no tool.
The data on candlestick patterns is very noisy. The same pattern under different names, on different timeframes, gives completely different results. And a person looking for them on a chart will always find them.
So treating patterns as an entry signal is weak. A better use is for timing — when you have already decided for some other reason that something is to be done here, the pattern can tell you whether to do it now or wait a little.
And an honest point that is rarely made: on small timeframes, candlestick patterns become almost useless, because every candle there looks like some pattern or other. They mean something only when enough trading has gone into a single candle.
Triangle, flag, wedge, channel — the names differ but these are all different forms of one thing: the part in the middle of a trend when price stops to rest.
Two things happen during that pause. Those who wanted to take profit quickly leave, and those who were left behind get their entry. When both jobs are done, price starts moving again, often in the same direction it was heading before.
This is why the success rate of continuation patterns is often better than that of reversal patterns — they are with the existing direction, not against it. And for the same reason the stops on them are small, because the area of consolidation is itself narrow.
Reversal patterns are those that signal the end of a trend — double top, head and shoulders, and the like. One thing is common to all of them: price tries to make one more high and fails.
The logic follows from the point on structure. An uptrend meant making higher highs. When one high does not form, that definition itself is breaking. The name of the pattern is just a form of that breaking.
The most common mistake with reversal patterns is taking the trade before it is complete. A head and shoulders that you can see is nothing until the neckline below breaks — and most of them never get completed. Trading a half-formed pattern is betting on something that has not happened yet.
With most patterns, a method for working out a target is given — usually adding the pattern's own height to the breakout price. This gives a measured move.
The logic behind it is that the width of the consolidation shows how that name was behaving at the time. A wide area points to a volatile period, and in such a period the next move is also expected to be large.
Its limit is that it is an estimate, not a promise. Many moves stop before that target and some go far beyond it. So making that number an exit rule often turns out costly — because the trades that pay the most are the ones that get cut at that number.
A better use is to look at that number before taking the trade, not to exit with it. If the measured move is much bigger than your stop distance, the trade's mathematics works. If it is about equal to the stop, there is no particular reason to take the trade, however clear the pattern is.
This is the one check that weeds out many clean-looking setups, and it needs no tool, only a comparison of two distances.
The most useful information often arrives when a clean pattern does not work. If everything was pointing one way and price went the other way, it means someone very large was on the other side who was not visible.
This is why experienced people treat failed patterns as a separate kind of setup. A broken breakout that comes back into the range often travels to the other end of the range — because those who came in on the breakout are now trapped and need to get out.
This understanding comes from the mindset where you see a pattern as a situation, not a prophecy. When the situation changes, the new situation has its own information, and it is often clearer than the old one.
Price tells you what happened. Volume tells you how many people agreed to it. A move that happened on very low volume belongs to few people, and it can reverse easily.
This is why the most useful application of volume is in comparison, not in absolute numbers. How much higher or lower is today's volume than its own average — that is the only question that matters.
And one thing that is rarely said: volume is dependable only when the name trades enough. In small names a single large order changes the whole day's volume, and any analysis of volume there is almost useless.
Rising volume with a breakout is considered a well-known confirmation, and its logic is sound — many people have to come in for a level to break.
The reverse is more useful. If a breakout happened on very low volume, the chance of it coming back is higher, because the force needed to break that level is simply not visible.
This can be used as a filter — do not take every breakout, take only those where participation shows. The cost is that you will miss some good moves that started quietly. Every filter has this cost, and the decision rests on which kind of mistake you want fewer of.
The biggest limit of volume is that it does not tell you who the buyer was. Price rising on increased volume looks bullish, but it could be one large seller who found buyers higher up.
The second limit is that delivery and intraday volume are different things. If a day's big volume was mostly intraday, it was entirely wiped out by evening and created no holding. Delivery data comes separately and can tell a different story.
So volume should not be treated as a standalone signal. It is a second opinion that strengthens or weakens what price is saying, and that is where it does its best work.
The ordinary volume chart shows volume over time — how much on which day. Another way is to split volume by price — how much trading happened at which price, no matter on which day.
This view directly tells you at which price the most people are stuck. Where the most trading happened, the most positions were built, and that price becomes the strongest level.
Its second benefit is to show the empty areas where very little trading happened. Price often passes through such areas quickly, because there is nobody there to stop it. It gives a good idea of where a move will be slow and where it will be fast.
Its limit is the same as that of every volume tool — it does not tell you whether there were buyers or sellers there. It only tells you there was a crowd. Which way the crowd will go is a separate question.
The entire logic of a breakout lies in stop orders. Above a resistance there are many stops lying in wait — those of people who are short, and the buy orders of people who want to buy if price goes up.
When price gets there, all of these trigger together. That is why the move after a breakout is fast — it is built not from new decisions but from orders that were already sitting there.
And that is why the price right after a breakout is often the worst entry. There you are buying after all the people whose orders have already been triggered, and the question of who the next buyer after you will be seems very far away at that moment.
A fakeout happens when price goes beyond the level and comes back. It is so common that many people treat its happening as normal and the breakout as the exception.
The reason is the same orders. For a large participant, the stops above that level are a pool of liquidity — there he finds the other side to fill his large order. Taking price there is a cost he bears.
There is no complete cure for it, but there are two half-cures. One — not going in immediately on the breakout, and waiting for a retest. The other — going in with small size and adding if the breakout holds. Both have a cost: some fast moves will slip out of your hands.
A retest is when price comes back to the level after the breakout and moves ahead again from there. It is better for entry because the stop can be kept close — just beyond that level.
Its downside is that not every breakout gives a retest. The fastest moves often run without coming back, and the one who waits for a retest is never in them.
So this is a pure trade-off and has no right answer. For those who want fewer trades and a better entry, the retest is fine. For those who do not want to miss any big move, going in on the breakout with small size is fine. Both paths work; trying to run both on one trade does not.
A moving average is an average that shifts forward with every new price. Its whole job is to reduce noise and show direction, and it does this job well.
Its basic limit is that it always trails behind. It is an average of old prices, so it shows new changes late. This is not a flaw, it is its definition — and that is exactly why it stays calm.
A longer-period average is calmer and turns later. A shorter-period one turns sooner but turns wrongly more often. These are two ends of one trade-off, and no setting escapes that trade-off.
Two averages crossing each other is the most taught signal, and it works fine in a trending market. In a range it works very badly, and the market spends most of its time in a range.
The result is that a crossover system makes its money from a few big trends in a year and takes small cuts the rest of the year. It is a sound approach if you can endure those cuts. Most people cannot endure them and abandon the system midway — right before the very period the whole system was built for.
So when choosing a crossover, the question is not whether it works. The question is whether you can stay through that long boring stretch where it gives nothing.
A common claim is that price takes support at a particular average. This shows up sometimes and has half a logic too — many people are watching that same average.
But there is no mechanism behind it. An average has no force of its own; it is only a calculation. When it seems to work, it is often because it happened to be near some real level at that time.
So rather than using an average as a level, it is better to use it as a direction filter — whether price is above it or below it, that is all. The answer to that one question is quite dependable, whereas expecting an exact bounce off it is not.
Which period is best — this question is asked a lot and has no answer, because the question itself is wrong. Each period shows the average of people from a different time span.
A short period shows the price of those who arrived recently. A long period shows those who have been sitting for months. Both are true, they are just about different people.
So the choice of period should match your position's holding time. If you hold positions for a few weeks, you should look at the average of those who have been sitting for a few weeks. That is one logic; the common numbers in use are just habit, not mathematics.
And one thing many people do not accept — the most widely used numbers work slightly better for exactly this reason, because many people are watching them. It is a self-fulfilling effect and it is real, but it is weak and cannot be relied on alone.
Momentum indicators measure the speed at which price is changing, not the level of price. They tell you whether the strength of a move is growing or fading.
Their entire calculation comes from price itself, so they carry no new information. They show the same information in a different form, and in that form some things become easy to see that are not noticeable on the price chart.
This is an important point many people do not accept. An indicator can never lead price, because it is built from price. Even any indicator called "leading" is really just a faster calculation.
This is the costliest misunderstanding. An indicator being in the overbought zone does not mean price will fall. It only means price has risen quickly.
In a strong trend an indicator can sit in that zone for weeks while price keeps rising. In such a period, "it's overbought so sell" is wrong almost every time, and that is the very period when the most money is made.
So a better use of momentum indicators is not to take trades against the trend, but to identify the resting phases inside the trend. In an uptrend, the indicator cooling off a little can be an entry opportunity — not a reversal.
Momentum tools come in two large families, and understanding the difference is enough.
The first family moves within a fixed range — meaning its number stays within a limit. Its advantage is that the numbers of different names can be compared, because everything is on the same scale. Its drawback is that once it reaches its limit it cannot say anything more, while price can go much further.
The second family is built on the gap between two averages and has no fixed range. Its advantage is that it also shows the level of strength — the wider the gap, the stronger the move. Its drawback is that its numbers cannot be compared across different names, because they depend on each name's own price.
Both come down to the same thing — whether speed is rising or falling. So applying both together does not add information, it only fills up the chart. Choosing one and understanding it properly is more useful.
Divergence is said to occur when price is making a new high but the indicator is not. It is roughly taken to mean that the strength of the move is fading.
There is some truth in it, but a big warning comes with it — divergence shows up very often and works very rarely. In a strong trend, divergence forms again and again and price still keeps going.
So treating divergence as a standalone signal turns out costly. Its job is that of a warning — pay attention, and if structure breaks too, then do something. Waiting for structure is what makes divergence useful.
Volatility bands draw a zone around price that widens and narrows with volatility. Their most useful application is the width, not the edge.
When the bands become very narrow, it means price movement has dropped. After such calm, a sharp move often comes — which way, the bands do not say, and that is their biggest limit.
Touching the edge is not a signal in itself. During a trending move, price can keep clinging to the upper edge. Trying to take the opposite side is the same mistake described in the overbought section.
The average range measure tells you how much a name usually moves in a day. It says nothing about direction, and that is its strength — it gives a number without an opinion.
Its most useful applications are in two places. One, in setting the stop distance — so that it sits outside the daily noise. Two, in setting target expectations — if a name usually moves two percent in a day, expecting five percent intraday will be wrong almost every day.
This second use is the least common and the most useful. Many people wait for targets that the name has never given in the whole of last year. Looking at one number fixes that whole expectation.
A gap forms when today's price opens away from yesterday's price, with no trading in between. It means something happened while the market was closed that changed opinion.
Roughly three kinds of gap are seen:
• A gap after a quiet range — this is often the start of a new phase.
• A gap in the middle of a running trend — this is often a sign that the trend will continue.
• A gap at the end of a long trend — this often comes around the end of that trend, when the last people arrive.
All three look alike and mean completely different things. The difference lies only in where it formed — which brings back the same point, that nothing means anything without context.
The first part of the day is different from the rest of the day and should be treated that way. The overnight news is absorbed into price in that period, and the range is widest then.
This means stops taken in that period are hit most easily, and levels from that period are the least dependable. A level formed in the first five minutes is almost no level at all — there was just a jerk there.
Many people therefore wait for the range of the first stretch to form and act on its break. The logic is that by then the effect of the overnight news has come into the price, and the move after that is built from new decisions.
The cost is the same as for every wait — sometimes the biggest move of the day happens in that period and you are not in it. This is a choice, not a rule, and it should fit your strategy.
A common saying is that every gap gets filled. This is a half-truth. Many gaps do get filled, but some do not for years, and the money sitting waiting for that would have been better placed elsewhere.
The practical use is to treat the other edge of the gap as a level, not as a prophecy. It is a place where a reaction can come — that is all.
And one important caution: the first hour of a gap day is almost always the most volatile. Decisions taken in that period are often the worst, because the price range is very wide at that time and any way of placing a stop turns out costly.
How a name is doing on its own is incomplete information. The full picture forms when it is compared with the market.
If the index is falling and a name is staying where it is, that name is actually strong even if its chart looks flat. And if the index is running and a name has only risen a little, it is weak even if it is in the green.
The simplest way to make this comparison is to look at both in percentage terms over the same period. It needs no indicator, and it is a small habit that greatly improves the decision of which name to pick.
A name moves with its sector, and the sector with the market. So it is useful to look at three levels — the market, the sector, and then the name.
In most big moves the sector's share is bigger than the name's own share. A good name in a weak sector often stays in the same place, and an ordinary name in a running sector moves ahead.
The most practical form of this is to check the sector before choosing a name. It is a one-minute job and often makes more difference than all the analysis done for hours on that one name.
An index is an average and it can deceive. If a few big names in that index are running, the index will look up even if most of the names inside it are falling.
The idea of breadth exists to see this — how many names are up and how many are down, or how many names are near their highs. It is a simple count and it tells you the health inside the index.
Some common breadth questions:
• How many names rose today and how many fell — what is the ratio between the two.
• How many names are above their long average — this is, roughly, a thermometer for the whole market.
• How many names are making new highs, and is that number rising or falling along with the index.
The most useful situation is when the index is making a new high but the names inside it are getting fewer. It means the rally has narrowed down to a few names. This is a warning, not a timing signal — this situation can last for months.
The reverse is just as useful. After a long fall, if the index is still going down but the number of falling names is shrinking, the selling is getting tired. Here too you get no timing, only context.
Eight indicators on a chart are not eight separate opinions. They are all made from one thing — price and volume. So their agreeing with each other is not confirmation.
Understanding this clears up a lot. If three momentum indicators are saying the same thing, they have done the same calculation three times, in slightly different ways. They are not proving each other right.
Real confirmation comes from a different kind of information — structure, volume, the sector's behaviour, or another asset. Piling on tools from the same family only raises confidence, not the probability of being right.
Every indicator has settings and they can be changed. Turning the settings on old data until the result looks good is the most common and best-hidden deception.
The problem is that the old data happened only once and many things in it were coincidences. Fitting settings to those coincidences produces a system that looks very good looking backwards and does nothing going forward.
A half-defence against this is not to keep the settings too precise. If your system works only when one particular number is used, and becomes useless on the numbers around it, then it is not a system — it is a coincidence in that data.
A backtest is a useful tool and has some very well-known lies. The biggest is that on an old chart both the entry and the exit look clean, while at the time a lot was going on between the two.
Common backtest lies:
• Not accounting for charges and slippage — which changes almost every result.
• Taking data only for names that exist today, leaving out those that disappeared along the way.
• Assuming liquidity that was not actually there at the time.
• Testing only on the period that suited your system.
This does not mean a backtest is useless. It means a backtest result is an upper limit, not an expectation. The real result will be lower — the only question is how much lower.
A system should be only as complex as it can be run every day without thinking. A system that demands an explanation on every trade is not a system, it is an opinion.
Most systems that work have very few rules. A direction filter, a setup, an entry, a stop, an exit, and a sizing rule. More rules than this were often added to fit past data.
A common sign that a system has become more complex than needed: if you cannot explain it to someone else in two minutes, you will not be able to run it properly yourself on a bad day.
Every system has three parts and all three should be written down:
• Entry — under what conditions to take a position. This matters least, although the most time goes into it.
• Exit — under what conditions to get out, both when wrong and when right. This matters more than entry.
• Size — how much to take. This matters most of the three and gets the least time.
Fixing this inverted ratio is often the biggest improvement. Two people running the same entry rules can get completely different results, just from the difference in exit and size.
Even after the system is written, one thing is missed — how a trade will run from start to end. That whole journey should be written down in advance, otherwise every step along the way becomes a new decision.
The steps in a trade's life, written in advance:
• Setup found — which condition being met brought this trade onto the list.
• How the entry will be made — at market, at a limit, or when price crosses some level.
• Where the stop is, and why it is there.
• What the size is — derived from the stop distance, not thought of separately.
• What happens if the trade goes right — when and under what condition the stop will move ahead.
• What the exit condition is — a price, a time, or a change in structure.
Of these six lines, the fifth is the most overlooked. Most people write the entry and the stop, and think about what to do when the trade goes right at that moment — when they are most excited and least objective.
And one line that is not in here but should be: what happens to this trade if a big gap comes tomorrow. The answer to it often makes the size smaller, and that is its whole purpose.
The most honest way of running a new system is to run it on forward data, not on past data. That means starting from today, at small size, and writing everything down.
It is slow and boring, which is why almost nobody does it. Its advantage is that everything a backtest leaves out comes in on its own — slippage, unfilled orders, and the biggest thing of all, that at the time you did not know what would happen next.
Only after at least a few months and enough trades does that data mean anything. Whatever shows before that is noise — good or bad.
After putting eight indicators on a chart, price itself stops being visible. And price was the one thing that held the real information.
The reason behind this is understandable — adding more tools makes us feel that we are seeing more. In reality we are looking at one thing eight times and dividing our attention over it.
A good test is this: if removing an indicator would not change any of your decisions, it is there only for comfort. And tools there for comfort are the costliest, because they give confidence without giving any information.
Hours of analysis give a satisfaction in themselves, and that begins to make you feel that something must now be done. This is a trap — the work done is no reason to take a trade.
The honest result of most analysis should be that nothing needs to be done right now. A large part of the market is like this. But a person who has watched charts for three hours cannot accept the answer "nothing", and ends up finding some setup or other.
This is why more effort often results in more trades, not better trades. And more trades mean more charges and more chances to make mistakes — both direct losses.
A small habit breaks this: at the end of the analysis, write that there is no trade today, when that is the answer. Writing that one line turns it into a decision, not a missed opportunity.
If a few good trades happen on one name, you form a relationship with it. After that you start seeing more setups on that name's chart, and that seeing is not real.
The reverse also happens — a name on which a big loss was taken gets abandoned, even if good setups keep appearing there. In both cases the decision is being made not from the chart but from memory.
Rebuilding the watchlist from scratch from time to time fixes this. And one honest question helps: if I did not know this name's history, would today's chart tell me to do anything?
Moving to a bigger timeframe when the position goes wrong and coming down to a smaller one when it goes right — these are two forms of one mistake, and it shows up in almost every new trader.
The result is that losses get longer and profits get shorter. The same asymmetry that breaks every system builds up from right here on its own.
The cure is to write down at entry — which timeframe this trade is on. That one line ends the whole argument later.
The human brain's job is to look for patterns, and it finds them even where there are none. Even on a random chart you will find triangles and head and shoulders.
There is only one way to avoid this — decide in advance what you are looking for. If your system runs on three setups, you look for only those three on the chart. The fourth thing you see is not part of your system, however clear it looks.
Drawing a level in advance is honest work. After price arrives, nudging that level a little further along — so that it still seems intact — is the best-hidden deception.
It happens almost automatically and the person does not even realise it. Support breaks, and right away a new support a little lower starts to appear. That new level was not on the chart before; it appeared because the old one broke.
There is only one cure — write the levels down before taking the position, on a sheet if not on the chart. A written level cannot shift. A level kept in the head always shifts in the direction your position is in.
Taking a trade after seeing someone else's chart does not work because a chart only tells you the entry. Their stop, their size, their timeframe and their exit condition — none of it is in that picture.
And when that trade starts going wrong, you have no basis of your own on which to decide whether to stay or exit. This is why copied trades get dropped the quickest and cost the most.
The gist of the whole article can be put in a short sequence, and every part of that sequence has appeared in detail somewhere above.
While looking at a chart, thinking in this order works:
• What is the structure on the larger timeframe — trend, or range.
• Are the market and the sector in the same direction, or is the name moving alone.
• Is there a meaningful level nearby — and why did that level form.
• What has volume said recently — has participation risen or fallen.
• If I am wrong, at what price will it be clear — and how far is the distance to that price.
• What size does that distance give me — and is the trade still worth taking.
The thing to notice is that there is no indicator anywhere in this sequence. Indicators go on top of this framework, they do not replace it. People who do the opposite start from indicators and never reach structure.
The second thing to notice is that the last two steps of this sequence are not about the chart at all. They are about risk. A chart tells you where something is happening; it cannot tell you how much you should put on it. And most accounts break in the second, not in the first.
The field of technical analysis is very large and most of it is repetition. If time is limited, a few things give more than all the rest.
Structure is at the top — it is the basis of almost every other tool and has no setting. After that, levels, because there is a real mechanism behind them. After that, volume, only as a comparison. Every tool after these three gives diminishing returns.
And one thing that is not in technical analysis at all, but without which it is useless — position size and the exit rule. An average setup with the right size gives a better result than an excellent setup with the wrong size. This is written nowhere on a chart, and it matters the most.
A common question is how much time to spend looking at charts. The answer is much less than most people expect.
For someone working on a swing timeframe, twenty to thirty minutes a day is enough — looking at the watchlist charts, writing down levels, and adjusting orders. Giving more time than that does not increase the analysis, it only increases the itch.
And when you give that time matters too. Time given after the market closes is the best, because price is not moving then and decisions are made with a calm mind. Time given during the market almost always produces a reaction, not a plan.
There is another side to this. People who sit in front of the screen all day often do so because they see trading as a job, where more hours give more results. Here that relationship is inverted, and that inversion is the strangest and most important thing about this work.
The biggest limit of technical analysis is the same as its strength — it looks only at price. It does not know how the business is, whether someone already knows that news, or whether some policy is about to change tomorrow.
So it can never give a complete answer. It is a very good tool for timing and risk, and a very bad tool for prophecy. For those who use it in the first way, it works for years. Those who use it in the second way keep running from one system to another, thinking they have not yet found the right settings.
This is why the most effective traders often use the fewest tools. Their charts look plain — price, a few levels, and perhaps one average. That plainness does not come from any lack; it comes from the years in which they removed, one by one, the things that changed no decision.
The right settings will never be found, because the problem was never in the settings. A chart cannot tell you what will happen. It can tell you what is happening now, and when you will find out if you are wrong — and for a trader, that second thing is more useful than the first.
This article is for educational purposes only — it is not a buy or sell recommendation, makes no claim of guaranteed returns, and is not advisory. Rules, rates and mechanics are changed from time to time by the exchanges, SEBI and RBI, so verify every number yourself with your broker or an official source.