ATR Indicator: How to Measure Volatility and Set Stops

Two traders set a 0.5% stop. The first trades BTC, the second got into a coin that listed yesterday and moves 4% a minute. The first trader's stop sits well outside the noise, the second one's gets taken out by the very first wick. The problem isn't discipline and it isn't risk size. The problem is that 0.5% means completely different things on different instruments, and the ATR indicator exists precisely to put that difference into numbers.

ATR converts a stop from abstract percentages into units of the actual movement a specific coin is making right now.

We'll go through the formula, the settings for crypto, position sizing, and how to tie the indicator's readings to the order book and the tape.

What the ATR Indicator Is

ATR (Average True Range) shows how far price travels on average within one bar of the chosen timeframe. Not where it's going, but how wide its steps are.

Welles Wilder came up with it in 1978 for commodity futures, where gaps and limit moves happened regularly. The same Wilder gave the market RSI, Parabolic SAR and ADX. Out of that set, ATR turned out to be the most durable, because it doesn't try to predict direction and doesn't give entry signals. It measures. Like a ruler.

The key difference from the usual percentage volatility: a standard calculation takes close to close, while ATR accounts for the entire distance covered inside the bar, including wicks and gaps between candles. For a scalper the difference matters, because what takes out a stop is the wick, not the closing price.

The Average True Range Formula

The calculation goes in two steps.

First, True Range (TR) is calculated for each bar. You take the largest of three values:

TR = max(

    High - Low,                   // range of the current bar

    |High - Close(previous)|,     // gap up plus the move

    |Low  - Close(previous)|      // gap down plus the move)

The three options exist because of gaps. If price closed at 100 and opened at 105, the candle's range can be tiny while the real move against the position is five points. The second and third options catch that.

Then TR is averaged using Wilder's method:

ATR(first value) = (TR1 + TR2 + ... + TRn) / n

ATR(current) = [ ATR(previous) × (n - 1) + TR(current) ] / n

Where n is the indicator period, 14 by default.

The second formula is exponential smoothing, not a simple average. Old values don't drop out of the calculation abruptly, they decay gradually. That's why ATR doesn't jump after every single outlier candle, but doesn't ignore it either.

A separate nuance for crypto. The market trades 24/7, the exchange doesn't close overnight, so there are almost no gaps and TR equals plain High − Low in 95% of cases. Gaps show up in three situations: at the listing of a new asset, during technical trading halts on the exchange, and during liquidation cascades, when the order book is empty and price jumps across dozens of levels without a single trade. It's exactly in those moments that the second and third lines of the formula start working and ATR expands sharply.

What ATR Shows

The indicator is plotted as a line in a separate pane below the chart. The value is read in the quote currency. ATR(14) = 95 on the BTC/USDT chart means the average bar over the last fourteen periods covered 95 dollars edge to edge.

The absolute number on its own says little. 95 dollars on BTC at a price of 67,000 is 0.14%. The same 95 dollars on an asset priced at 300 dollars would be a disaster.

That's why a normalized value is used to compare coins against each other:

ATR% = (ATR / current price) × 100

Now that's a number you can work with. It answers the question of which coin is moving today and which one is standing still.

The spread between BTC and a first-day coin is 26x. A stop considered reasonable on bitcoin will trigger within the first twenty seconds on a listing.

The second thing ATR shows is the market regime. A rising line means the range is expanding, participants are active, the tape is speeding up. A falling or flat line means compression, the market has gone into accumulation, there are no moves.

An important detail beginners miss. ATR has no direction. It rises the same way on a flush down and on a squeeze up. The indicator answers "how much," never "which way."

ATR vs. Other Ways to Measure Volatility

Average True Range isn't the only tool for gauging price range. It's useful to know how it differs from its neighbors on the shelf, because the question of which volatility indicator to pick comes up for every beginner.

The difference between the last two rows deserves its own paragraph. A coin with a 0.3% daily change looks dead in the screener, but if it went 2% back and forth fifteen times during the day, ATR sees that and 24h change doesn't. A scalper gets paid by the range, not by the net price shift.

By the way, Bollinger Bands are built on a similar idea of measuring dispersion, only through standard deviation from a moving average, and trading off their boundaries makes sense precisely when ATR is high.

If the basic theory is still hard to absorb, watch the free lesson 4 from our course for beginners. It breaks down how professionals read the market through the order book and clusters, and it sits well on top of this chapter.

How to Use ATR in Trading

Three practical applications, from the most common to the most underrated. Stop loss, position sizing, and the "we're not trading today" filter.

ATR-Based Stop (ATR × 1.5 / ATR × 2)

The base logic is simple. A stop has to sit outside normal market noise, otherwise it gets knocked out by a random fluctuation rather than by the idea being wrong.

Long stop  = Entry price - (ATR × multiplier)

Short stop = Entry price + (ATR × multiplier)

The multiplier is chosen to fit the style and the timeframe.

The 1.5 multiplier became an unofficial standard for a reason. Statistically, a pullback inside a working move rarely exceeds one and a half average bar ranges, while two or more is usually a break of structure.

I normally take ATR × 1.5 as the base, then look at the order book and move the stop to the nearest density level sitting beyond that calculated point. A density level, meaning a large cluster of limit orders at a single price, works as a physical barrier for price. Bare math gives you the distance, the order book gives you a specific level.

A separate word on trailing. The classic construction is called the Chandelier Exit, where the stop hangs below the high of the move at a distance of three ATRs:

Trailing stop (long) = Highest high over N bars - (ATR × 3)

The stop is only pulled up and never moves down. In the terminal you do that by dragging the stop line right on the chart with the mouse, or with a hotkey that moves it to breakeven instantly.

What you absolutely must not do is drag the stop into the red when price goes against you. Moving the stop loss consistently stays at the top of the beginner mistake list, and the consequence is always the same: a growing loss, all the way to liquidation. ATR gives you the distance before the entry, not after it.

For more on the logic of placing protective orders, read the article "Take Profit and Stop Loss: How to Set Them Correctly".

Position Sizing Through ATR

This is where ATR really opens up, and this is what a minority actually uses.

The usual approach is a fixed size. A trader works with lots of $1,000, $2,000, $5,000. Convenient for speed, but risk floats around. On a quiet bitcoin, a 0.14% stop gives a $1.4 loss per thousand; on a volatile altcoin the same setup gives $35.

The correct logic flips the calculation. First you fix the risk amount, then you size the position from it:

Position size (in coins) = Risk in $ / (ATR × multiplier)

Let's run the numbers. A $10,000 account, 0.5% risk per trade, so a maximum loss of $50.

BTC/USDT. ATR(14) on 5m = 95, multiplier 1.5, stop distance 143 dollars. Size = 50 / 143 = 0.35 BTC. At a price of 67,200 that's a $23,520 position, meaning roughly 2.3x leverage.

SOL/USDT. ATR(14) on 5m = 0.62, multiplier 1.5, distance 0.93. Size = 50 / 0.93 = 53.7 SOL. At a price of 178.40 that's $9,580, leverage slightly under one.

The positions differ 2.5x in notional, while the dollar risk is identical. Exactly $50 in both cases. More volatile asset, smaller size. Calmer asset, larger size.

What this gives you over time. A run of ten stops in a row costs the same whether you traded bitcoin, Solana, or a fresh listing. Your equity curve stops depending on which coins happened to show up this week. That's what systematic trading actually is, not the number of indicators on the screen.

The basics of capital allocation are covered in detail in "Money Management in Trading: Managing Your Capital".

Volatility Filter: When Not to Trade

Scalping works on movement. No movement, no profit — but the fees and the spread don't go anywhere.

Let's do the entry math. The taker fee on futures averages 0.05%, entry and exit make up 0.1% of turnover. Plus the spread. On a liquid instrument with $300 million in daily turnover the spread is around 0.011%, which is almost invisible, but on an illiquid one it balloons to fractions of a percent. Binance, Bybit and OKX have base fee tiers that barely differ, while on small venues the fee difference is capable of killing your whole statistics.

Hence a simple cutoff rule. If ATR on your working timeframe is less than four round-trip fees (entry plus exit), there's nothing to trade. At 0.1% per round trip, that's a threshold of roughly 0.4% of price on the timeframe where you hold the trade.

The practical filter setup looks like this:

• Calculate the average ATR over the last 20 periods and compare it with the current value
• Current ATR below 60% of the average: the market is asleep, the tape is standing still, any entry is a lottery played against fees
• ATR in the 80-150% band of the average: normal regime, standard multipliers work
• ATR above 200%: a news-driven squeeze or a liquidation cascade — you can trade, but cut size in half, because a stop based on the old calculation is already inadequate
• ATR above 300% with an empty order book: don't get in at all, slippage will eat more than the move gives

The second use case is picking coins before a session. The working criteria for a scalping instrument are strict: price change of more than 10-15% over 24 hours, daily volume above $100-150M, more than 800 thousand trades. ATR% works as a fast numerical cut on the first criterion. Run the list through the screener, sort by ATR%, take the top five rows with adequate volume.

The reverse situation happens too. A coin shows ATR% near 5%, but the trade count is 40 thousand. That's not volatility, that's a ragged chart on an illiquid instrument where two trades move price by a percent. Filter that out immediately.

For a fuller treatment of the nature of price swings, read "Cryptocurrency Volatility: How to Use It".

ATR Settings for Crypto

The default period of 14 came from the seventies, when Wilder worked with daily bars on commodity futures. Crypto trades around the clock, and a one-minute bar here carries more events than a daily corn bar did in 1978. The settings are worth revisiting.

Period 14 vs 7 vs 21

The period sets the indicator's memory depth. The shorter it is, the faster the reaction and the more jitter.

In practice it's fine to run two ATRs at once. A short period of 7 on your working timeframe shows what's happening right now. A period of 21 on the hourly shows the backdrop.

The divergence between them is informative on its own. ATR(7) on the one-minute has tripled while ATR(21) on the hourly is flat: that's a local spike, most likely a stop hunt that will fade within a couple of bars. Both rising in sync: a real range expansion has started, the market regime has changed.

I tested this pairing on BTC/USDT and on the top altcoins, where it reads most consistently. On coins outside the top three hundred by turnover the signal is noisy.

ATR Across Timeframes

Volatility scales roughly as the square root of time. That follows from the statistical nature of a price series, and in crypto the relationship holds fairly accurately on liquid pairs.

ATR(timeframe 2) ≈ ATR(timeframe 1) × √(T2 / T1)

The practical meaning: knowing ATR on the five-minute, you can estimate ATR on the hourly without switching charts.

Why this matters in actual work. You're scalping on the one-minute chart with a 60-point stop, while the hourly ATR reads 330. That means your stop distance fits inside a single hourly bar five times. One wide hourly candle will take you out with high probability, even if you got the direction right.

The rule that follows. Your stop on the working timeframe has to be proportional to the ATR of the timeframe on which you make the decision, not the one you use to find the entry. Decision made on 15m, take ATR(15m). You look for the entry point in the order book and on the one-minute — that's a question of execution, not stop size.

Deviations from the square-root relationship also tell you something. If ATR(1h) is well above the value implied by ATR(5m), there's a directional trend inside the hour, moves are adding up rather than canceling out. If it's below, the market is chopping in a range, and the hourly candle comes out short despite active internal churn.

ATR, the Order Book, and the Tape

This is where the part that makes the indicator worth understanding at a professional level begins.

ATR looks backward. It's built on closed bars and answers the question of how the market moved over the last N periods. The order book looks forward. It shows limit orders sitting right now, not yet filled. One tool gives you distance, the other gives you a level.

Separately, each works at half capacity. A stop placed on ATR alone often ends up in empty space between levels. A stop placed only behind a density level sometimes sits two ticks from the entry and gets knocked out by any twitch.

The working combination looks like this. Calculate the distance as ATR × 1.5. Look at which density levels sit beyond that mark. Place the stop behind the first such level plus a few ticks of buffer. If there's nothing beyond the calculated distance for another two ATRs and the order book is empty, it's better to skip the trade, because there'll be nothing to protect the position with.

The Secret Terminal toolset is built directly for this task:

• Ruler (key R). Measures distance in points or percent right inside the order book window. You estimate ATR on the chart, carry the value into the book, and see how many price levels it covers and which orders fall inside
• Auto-stop. Places the stop at a set distance from the entry point immediately after the order fills. You enter the distance once based on the ATR calculation for the current coin, and from there it happens on its own
• Key C. Automatic adjustment of the order book preset to the coin's current volatility. Essentially the same ATR approach, only applied to the display rather than to the stop
• P, L, O filters with the mouse wheel. Dynamic adjustment of volume and cluster thresholds. A cluster, meaning the trade volume that went through at a specific price level over a period, needs a higher threshold when ATR has doubled, otherwise the order book turns into a Christmas garland
• Density map. Shows limit orders that have been sitting in the book for more than 30 minutes, across all coins at once. A ready-made list of levels you can anchor a stop to after calculating the distance
• Quotes module. Sorting by daily change and volume — the first sieve before the ATR% calculation
• Footprint chart. Shows exactly where inside the bar the volume went through, and explains why that bar's TR came out so wide

One more layer is the tape. The tape, meaning the real-time stream of all executed trades, answers a question ATR can't: why the range expanded. The tape accelerating with large prints in one direction means real money coming in. ATR expanding with a sluggish tape means the book is spreading out — the market maker pulled the orders and price is falling into a void on laughable volume. The second case is more dangerous, because a calculated stop there fills with slippage several times larger than expected.

The third element of the combination is clusters. ATR said the bar was wide. The tape said buying was aggressive. The footprint chart shows exactly which level inside the bar absorbed the bulk of the volume, and that's the level to anchor a stop to, not a round number.

Common Mistakes When Working with ATR

The indicator is simple, but ruining your statistics with it is easy. Four situations come up more often than the rest.

Mistake 1. One ATR for the whole trading day. A trader calculates in the morning that ATR on the coin is 0.0034, enters that distance into the auto-stop, and trades that way until evening. By midday a news item hits, the range triples, and the stop stays the same. Now every trade gets knocked out by noise. You need to recalculate before every entry — it's thirty seconds of work.

Mistake 2. Fitting the multiplier to the size you want. You want to enter with a bigger lot, so the multiplier magically shrinks from 1.5 to 0.8. Formally the calculation was done; in practice the stop was placed inside the noise. The logic works the other way around here: the multiplier is fixed by trading style, and size is fitted to it.

Mistake 3. Using ATR as an entry signal. The indicator line turns up and the trader reads it as "the move has started, time to jump in." There is no direction in that signal at all. Range expansion looks identical ahead of a run up and ahead of a flush down. More than one person has blown an account on this logic.

Mistake 4. Calculating with ATR without looking at the order book. The math produced a distance of 143 points and the stop went exactly there. But it's empty there, the nearest density level is 60 points further out, and price runs through that zone without stopping. A stop in empty space is a stop that fills worse than calculated. The second case above is exactly about this.

Mistake 5. Ignoring the higher timeframe. The stop is calculated from the one-minute ATR, while the entry decision was made on the fifteen-minute. The distance ends up several times smaller than a normal move on the scale where the idea lives. The trade dies before it gets a chance to start.

If you're just starting out and half the terms above still sound unfamiliar, there's a free crypto trading and scalping course on our YouTube channel: "Trading from Scratch | Free Crypto Trading and Scalping Course." Five lessons, from the basics of futures to breaking down the order book lesson 5. The course is free, no registration required anywhere.

When ATR Doesn't Work: Limitations of the Indicator

The indicator isn't perfect, and understanding its weak spots is more useful than knowing the formula by heart.

First. ATR lags by definition. It's calculated on closed bars, which means it reacts to a regime change with a delay of several periods. At the moment a news item drops or a liquidation cascade starts, real volatility has already risen fivefold while the indicator line has only just begun to curve up. For a scalper that's a serious problem, which is why pairing it with the tape is mandatory.

Second. It knows nothing about the reason for the move. Range expansion ahead of a funding rate payment (the periodic payment between longs and shorts on perpetual futures), on news, and on a technical stop hunt all look identical on the indicator, while these situations need to be traded differently.

Third. On illiquid instruments the calculation lies. If eight trades went through inside a five-minute bar, its High and Low were formed by random orders, and TR reflects the absence of liquidity rather than volatility. Formally ATR% will be high. In practice you won't be able to enter at the calculated price and exit at the calculated one.

Fourth. After a listing, a coin has no history. ATR(14) on the one-minute requires fourteen bars, meaning fourteen minutes, while the actual move happens in the first two or three. For trading listings the indicator is useless; other mechanics work there: spread width, tape speed, order book depth.

Fifth, the psychological one. A wide ATR-based stop creates a feeling of protection, which leads a trader to allow himself sloppier entries. Stop distance doesn't cancel the requirement for a quality entry point, it only protects against noise.

FAQ

• What does the ATR indicator show in simple terms?

  The average price range of one bar on the chosen timeframe, in the quote currency. ATR(14) = 95 on BTC/USDT means that over the last 14 bars the average candle traveled 95 dollars from low to high. ATR shows no direction whatsoever, only amplitude. That's why it's used to calculate distance, not to find entries.

• Which ATR multiplier should you use for a stop loss?

  For intraday trading the working range is 1.5-2.0. Scalping off density levels in the order book allows 1.0-1.2, if the entry point is dialed in to the tick. A multiplier of 3 and above is already swing territory; for intraday work such a stop eats the entire profit in one trade. The multiplier is fixed by style and doesn't change just to increase size.

• Which ATR period is better for crypto?

  Period 7 on the one-minute and five-minute charts to gauge the current moment, period 14 on 5m and 15m to calculate the stop, period 21 on the hourly to understand the backdrop. Running two indicators with different periods is more useful than hunting for one perfect setting. The divergence between the short and long ATR gives you information about the nature of the spike on its own.

• Can you determine the direction of a move from ATR?

  No. The indicator measures amplitude only and rises the same way on a flush down and on a squeeze up. Direction requires other arguments: the structure of levels on the chart, delta in the footprint, an imbalance in order book density, the flow of the tape. ATR answers "how much"; direction has to be sourced separately.

• How does ATR differ from percentage volatility?

  Standard volatility is calculated on closing prices and doesn't see the intrabar range. ATR takes the maximum of three ranges, including gaps between candles, so it accounts for wicks. And what takes out a stop is the wick, not the close. The gap between the two approaches on a volatile altcoin reaches a factor of one and a half.

• Does ATR work on first-day listings?

  Practically no. The indicator needs at least N closed bars to calculate, while the main move on a listing fits into the first few minutes. There you go by spread width, tape speed, and whether there are orders in the book at all. ATR becomes informative roughly an hour into trading.

• How often should you recalculate the ATR stop?

  Before every entry. Volatility on altcoins changes several times over within a few hours, and a position size calculated in the morning gives a completely different dollar risk by evening. One value for the whole day is the single most common mistake when working with the indicator. Recalculating takes under a minute.

What to Take Away

ATR gives no entry signals and doesn't replace order book analysis. It does one thing, but it does it precisely. It converts abstract volatility into a number from which stop distance and position size are calculated.

The workflow fits into four steps. Look at ATR on the timeframe where the decision is made. Multiply by the multiplier that fits your style. Move the resulting point out to the nearest density level in the order book. Size the position from a fixed risk amount, not from your habitual lot.

If you apply only the last point from this entire article, your equity curve will smooth out within a month.

Measure volatility in the same place where you trade.

Secret Terminal brings the chart with indicators, the order book with the density map, and the tape into one window. The ruler (R) measures ATR distance right in the order book, the auto-stop places protection at a set distance immediately after entry, and key C adjusts the order book preset to the coin's current volatility. No switching between tabs at the moment the tape takes off.