ATR and volatility stops
The answer
ATR, the Average True Range, tells you how far an instrument normally moves in one bar, including any gap from the previous close. It has no direction and it is not a signal.
It is the most useful number in this cluster, because it converts an arbitrary stop distance into one scaled to how much the thing actually moves — and because it lets you size every position so that being wrong costs the same amount every time.
Why this costs you money
Open your positions right now and look at where your stops are. Most readers will find at least 1 stop placed at a round number, or at a level chosen because "10% felt right", or at the price where the loss reaches an amount they do not want to exceed.
None of those has anything to do with the stock.
Here is what happens. You buy a stock that normally moves 3.5% in a day. You put your stop 2.5% below your entry, because 2.5% is a loss you are comfortable with.
Your stop is inside 1 normal day of movement. It will be hit. Not because you were wrong about the company, not because the trend failed, but because the stock did on Tuesday what it does most Tuesdays. You will be removed, and then you will watch it go where you thought it would go.
This is the single most common expensive mistake in retail trading, and it is invisible because each individual instance looks like bad luck. It is not bad luck. It is a stop placed by reference to your feelings instead of by reference to the instrument.
The mirror mistake costs more per event. You buy a stock that normally moves 0.9% a day and place a stop 10% away, because 10% is your standard. You are now risking 11 normal days of movement on one position. When it does go wrong, it goes wrong by an amount you never sized for.
A stop distance chosen without knowing the instrument's normal daily movement is a random number. Everything downstream of it — your position size, your risk per trade, your expected outcome — is built on that random number.
How it works
ATR is built in 2 steps, and the first step is the clever one.
Step 1. True Range
For each bar, take the largest of these 3 distances:
- Today's high minus today's low
- Today's high minus yesterday's close, as a positive number
- Today's low minus yesterday's close, as a positive number
Why 3 measures instead of just the high minus the low? Because of gaps. If a stock closes at 100 and opens at 88 the next morning, then trades between 88 and 90 all day, the high minus low is 2. The stock moved 12. True Range captures that, because measure 3 gives 12.
J. Welles Wilder Jr. introduced True Range in New Concepts in Technical Trading Systems in 1978, and the reason matters enormously for Indian readers. Wilder traded commodity futures, which had daily limit moves — a market could be locked at its limit with almost no trading, printing a tiny high-to-low range on a day when the market had moved violently. High minus low was useless there. True Range was designed specifically to see through that.
Indian equities have exactly that problem, in the form of price bands. Wilder's fix is 47 years old and was built for the situation Indian traders face every week.
Step 2. Average it
ATR is the average of the True Range over the last N bars. Wilder's default is 14, and he used a smoothing method that weights the recent past more than a simple average would. Most platforms use his method by default.
The output is a price, not a percentage. If a stock trades at ₹1,240 and its 14 day ATR is ₹31, that stock moves about ₹31 in a typical day.
Step 3, which is not in the indicator: make it comparable
Divide ATR by the current price.
₹31 ÷ ₹1,240 = 0.025, or 2.5% per day.
Now you can compare 2 completely different instruments. A stock with a 2.5% daily ATR and one with a 0.7% daily ATR are different animals, and no single stop distance or position size can be correct for both.
What it tells you, and what it does not
It tells you the size of a normal move. That is a genuinely new fact that price alone does not present clearly, and it is the basis of every sensible stop and every sensible position size.
It tells you when the character of an instrument has changed. ATR rising from 1.2% to 3.4% over 3 weeks means the market you are trading is not the market you started in, even if the price is the same.
It does not tell you direction. ATR rises in crashes and rises in violent advances. A high ATR is not bearish.
It does not tell you risk is low when it is low. This is the most dangerous misreading, and it deserves a full sentence: volatility is lowest immediately before it becomes highest. Quiet periods precede violent ones with striking regularity. A low ATR means the recent past was calm. It says nothing about tomorrow, and if you size your position from a low ATR without thinking about what is scheduled next week, you have used a calm measurement to take a large position into an event.
It does not look forward. ATR is an average of what already happened. Before a results announcement, a budget, or a central bank decision, ATR will be at its quietest and the real risk will be at its highest.
And it is not a signal. There is no "ATR crossover". If somebody shows you an entry rule based on ATR alone, they have taken a risk tool and dressed it as a prediction tool.
The decision rule
This is the part to keep. It is 3 lines and it changes how every trade is constructed.
1. Stop distance. Place your stop at least 2 × ATR from your entry, in the direction that would prove you wrong. Use 2 for shorter holding periods and 3 for positions you intend to hold for weeks. Never less than 1.5.
2. Risk per trade. Decide, once, what fraction of your account you are willing to lose on any single trade. For most people this is between 0.5% and 2%. Write it down. It does not change from trade to trade.
3. Position size. Now the size is not a choice. It is a calculation:
Number of shares = (Account value × risk %) ÷ (ATR multiple × ATR in rupees)
Work through what that does. Suppose your account is ₹5,00,000 and you risk 1%, which is ₹5,000.
- A stock at ₹1,240 with an ATR of ₹31, using a 2 ATR stop: your stop is ₹62 away.
₹5,000 ÷ ₹62 = 80 shares, a position of about ₹99,000. - A stock at ₹1,240 with an ATR of ₹9, using a 2 ATR stop: your stop is ₹18 away.
₹5,000 ÷ ₹18 = 277 shares, a position of about ₹3,44,000.
Same account, same risk, same share price, positions 3.5 times different in size. That is the correct answer. The quiet stock gets more capital because it takes more capital to lose ₹5,000 on it.
This is the whole point of ATR, and it is why this article insists it is a risk tool. It does not tell you what to buy. It tells you how much, and where to admit you were wrong, in units the instrument itself defines.
Try this now
Five minutes. Do this on a position you currently hold, with real money in it.
- Open the daily chart of a stock you hold. Add the ATR indicator with the default period of 14. Read today's value. Write it down in rupees or dollars.
- Divide it by the current price. That is the stock's normal daily move as a percentage. Write that down too.
- Now find your actual stop. If it is an order sitting with your broker, read the price. If it is a level in your head, write down the number you would actually sell at. If you have neither, write down "none" and read the paragraph below.
- Measure the distance from the current price to your stop, in rupees or dollars.
- Divide that distance by the ATR. Write down the result.
What you should see. A number. That number is how many normal days of movement your stop can absorb.
Most readers doing this for the first time get a number below 1. Some get 0.4. If your number is below 1, your stop is inside a single ordinary day's movement, and the ordinary behaviour of the stock will remove you from the position. You are not managing risk. You are guaranteeing an exit at a randomly chosen moment.
If your number is between 1 and 1.5, you are still inside the noise on any slightly active day.
If your number is above 4, ask a different question: is the position small enough that a move to your stop is a loss you planned for? A wide stop is only sensible with a small position.
Now do the last step, which is the one that changes behaviour. Repeat it for every position you hold. Write the ATR multiple for each one in a column. You will almost certainly find that your positions risk wildly different amounts — one might risk 0.3 ATR and another 5 ATR, with similar amounts of money in each. That column is the real risk profile of your portfolio, and until you have written it out you have not seen it.
Three real cases
1. J. Welles Wilder Jr., 1978 — the problem the formula was built to solve Wilder introduced True Range, ATR, ADX and Parabolic SAR in a single 1978 book. He defined True Range using 3 measures rather than the simple high-to-low range because the commodity markets he traded had limit moves: sessions where the price hit an exchange-imposed boundary and trading effectively stopped. On such a day the high-to-low range is tiny and the actual movement is large. Wilder's solution was to compare today's extremes to yesterday's close. This is the direct ancestor of the correction Indian traders need for price bands, and it was solved before most of them were born.
2. The Turtle traders, from 1983 — ATR as the unit of position size Richard Dennis and William Eckhardt trained a group of new traders in Chicago beginning in 1983, using a fully specified rule set. Their system used a volatility measure they called N, which was a 20 day average true range. Two things were defined in units of N: the stop was placed 2N from entry, and the position size was set so that a 1N move equalled a fixed small percentage of the account. The significance is not that the system made money. It is that a professionally designed system used volatility as the unit of account for both stops and size, which is exactly the discipline described in the decision rule above.
3. Adani Enterprises, January and February 2023 — where ATR understates the danger Following the publication of a short-seller report on 24 January 2023, shares in several Adani group companies fell heavily over the following weeks. Adani Enterprises repeatedly hit lower circuit limits during that period. On a circuit day, the recorded range is capped by the band. True Range does capture the gap from the previous close, which is exactly Wilder's fix working as designed — but when a stock opens at its limit and stays there, even the gap is capped by the band. Several consecutive band-limited days produce an ATR that is lower than the real movement. A trader sizing a position from that ATR would have taken a larger position than the situation warranted, at the worst possible moment.
The question that resolves it
A novice looks at a stop and asks: how much money will I lose if this is hit?
An expert looks at the same stop and asks: how many normal days of movement does this stop allow?
The first question is about the trader. The second is about the instrument, and only the second predicts whether the stop will be hit by noise.
There is a second question underneath it, and it is the one that separates a trader with a process from one without. Given that stop distance, how many shares should I own? If the answer to that is "as many as I felt like buying", then the stop was decoration.
What would make this wrong
If ATR-scaled stops were no better than fixed-percentage stops, then a portfolio using 2 ATR stops and one using a fixed 5% stop would be removed from positions at similar rates and with similar outcomes. Anyone can test this on their own trade history. Go back through your last 20 closed trades. For each one, look up the ATR at the time of entry and calculate how many ATRs away your stop was. Then check which of your losing trades subsequently recovered to your original target. If most of your stopped-out trades had stops inside 1.5 ATR and later recovered, you have measured the effect directly, in your own account.
The honest limits are 4, and they matter.
First, ATR is backward-looking. It tells you about the last 14 bars. Before a scheduled event, the last 14 bars are the worst possible guide. Widen manually before results, budgets and policy decisions, or do not hold through them.
Second, low ATR does not mean low risk. Volatility clusters and mean-reverts. The quietest readings often come just before the largest moves. Sizing up because ATR is low is how people take their largest position into their largest loss.
Third, an ATR stop does not protect you across a gap. Your stop is a price. If the market opens beyond it, you exit at the open. ATR sizing controls your loss on an ordinary adverse move. It does not cap your loss on an event.
Fourth, this article's stop multiples of 2 and 3 are conventions, not findings. They come from the trading literature and they are reasonable starting points. The principle — scale the stop to the instrument's own movement — is what matters. The specific multiple is yours to test.
In India
ATR is available on every Indian broker charting platform. It is used most often as the engine inside Supertrend rather than on its own, which is a shame, because the standalone use described in this article is worth more.
Three Indian features matter.
Price bands truncate the input. This is the central Indian issue with ATR. Most cash-market stocks have daily bands of 5%, 10% or 20%. On a day the stock is locked at a band, the True Range is capped at approximately the band width. Feed several such days into a 14 day average and the ATR falls. A trader using that ATR places a tighter stop and takes a larger position, at exactly the moment the stock has demonstrated it is capable of moving further than the exchange will allow in one session.
The practical correction: before using ATR on an Indian stock, look at the last 20 daily candles. If any closed at or very near a circuit limit, do not use the computed ATR for sizing. Use the band width itself, or the stock's ATR from before the event, or skip the trade.
Index versus stocks. NIFTY 50 and BANK NIFTY have no equivalent daily band, so index ATR is not distorted this way. Index-level circuit breakers exist at 10%, 15% and 20% and halt the whole market, but they are rare.
India VIX. India VIX is an index of expected 30 day volatility derived from NIFTY 50 index option prices. Unlike ATR, it looks forward, because option prices contain what people are willing to pay for protection over the coming month. A rising India VIX with a flat ATR is the clearest available warning that the market expects the near future to be more violent than the recent past. Reading the 2 together is one of the few genuinely independent pairings available to a retail trader.
In the United States
ATR is standard on all US platforms and is used directly for stop placement and sizing by a large number of systematic traders.
Three US features matter.
No cash-market daily bands. US equities have Limit Up-Limit Down bands that pause trading briefly around rapid moves but do not cap the day. A US daily range is therefore a fairer record, and ATR computed from it is a more honest number than the same computation on a band-limited Indian stock.
Overnight gaps are the dominant risk. Most US companies report results outside market hours. A very large share of the total movement in individual US stocks happens between the close and the next open. True Range includes those gaps, which is why ATR on a US stock with a history of large earnings gaps is meaningfully higher than its intraday movement would suggest. That is the indicator working correctly. It is also a warning that the ATR-sized stop will not be available on the day it is most needed.
The CBOE VIX. VIX measures expected 30 day volatility implied by S&P 500 index option prices. It closed at 82.69 on 16 March 2020, among its highest closes on record. Like India VIX it looks forward. The same pairing applies: a rising VIX against a flat ATR says the market expects more movement than it has recently had.
Where they differ, and what that tells you
Three real differences, each with a practical consequence.
1. India censors the range; the US does not. An Indian stock at a circuit limit prints a small range on a violent day. Wilder built True Range for exactly this situation, and it helps — it captures the gap to the previous close — but when the open is itself at the band, the whole day is capped and even True Range is truncated. So on Indian charts, ATR is a floor on real movement during stressed periods, not an estimate of it.
2. The US concentrates risk into the overnight gap. An American stock's ATR is often dominated by moves that happened when the market was closed. So on US charts, ATR is an honest measure of total movement, but a poor guide to what you can actually exit at. An ATR-based stop on a US stock through an earnings date is a number, not a protection.
3. The 2 volatility indices are not the same measurement. Both India VIX and CBOE VIX express expected 30 day volatility from index option prices, and India VIX uses a methodology derived from the CBOE approach. But the underlyings are different in a way that changes what the number means.
- CBOE VIX is built on options over the S&P 500, an index of 500 companies across the US economy.
- India VIX is built on options over the NIFTY 50, an index of 50 companies in which a small number of names carry a very large share of the weight.
A 50 name, concentrated index carries more single-company and single-sector risk inside its own volatility number. Indian index option activity is also concentrated around short-dated weekly expiries, which affects the option prices the calculation reads.
What that tells you is a translation rule. Do not compare the levels of the 2 indices as though they measure the same thing. A reading of 20 on India VIX and 20 on CBOE VIX do not describe the same amount of underlying uncertainty, because the things they are uncertain about are different portfolios. Compare each index to its own history instead. "India VIX is at the 90th percentile of its last 3 years" is a statement that means something. "India VIX is higher than CBOE VIX" is not.
Carry this
- ATR is how far this instrument normally moves in a day, gaps included.
- A stop closer than 1.5 ATR will be hit by ordinary noise, not by being wrong.
- Fix your risk per trade, then let ATR decide your position size. The size is a calculation, not a preference.
- Low ATR is not low risk. It is a description of a quiet past.