ATR Indicator Guide: The Most Useful Number in Risk Management

How ATR is calculated

True Range for each bar = the greatest of:
   (1) High - Low
   (2) |High - Previous Close|
   (3) |Low  - Previous Close|

ATR = a smoothed average of True Range over N periods
      (usually Wilder smoothing, similar to an EMA)

Why (2) and (3) matter:
   Previous close 100.  Today opens at 94 and trades 93 to 95.
   High - Low          = 2.00   (understates the move)
   |Low - prev close|  = 7.00   (captures the gap)
   True Range          = 7.00

A stop sized on high-minus-low alone would ignore gap risk
entirely, which is precisely the risk that matters most.

Using ATR for stops

A percentage stop treats all instruments identically, which guarantees it is wrong for most of them. An ATR stop places the exit at a distance proportional to how much the instrument actually moves.

Each stop is 2 ATR away, and the percentage distances range from 1.1 to 6.6 percent. A single percentage rule would be simultaneously too tight for some and far too wide for others, producing inconsistent risk and inconsistent behaviour.

ATR position sizing

Position size = (Equity x Risk%) / (ATR multiple x ATR x value per point)

Equity 80,000, risk 0.5% = 400 USD, stop = 2.5 ATR

Stock at 46.00, ATR = 1.15:
   Stop distance = 2.875 per share
   Shares = 400 / 2.875 = 139  (6,394 USD position)

Micro gold future, ATR = 22 points, 10 USD per point:
   Stop distance = 55 points = 550 USD per contract
   Contracts = 400 / 550 = 0.73  ->  0 contracts at this risk level

Both positions are evaluated on identical terms: what does a
normal adverse move cost, and how many units make that equal
to my risk budget?

The second case is instructive. ATR sizing does not tell you to take the trade with a smaller stop; it tells you the trade does not fit this account at this risk level. That honesty is the point of the method.

Other uses for ATR

• Trailing stops. The chandelier exit trails at the highest high since entry minus a multiple of ATR, adapting automatically as volatility changes. See trailing stops.
• Volatility filters. Comparing ATR(10) to ATR(50) identifies compression and expansion. Low ratios indicate a quiet market, which favours breakout setups and penalises trend-following entries.
• Profit targets. Setting targets in ATR units makes them comparable across instruments and adapts them to conditions.
• Normalising for comparison. ATR divided by price converts to a percentage, allowing volatility comparison across instruments at different price levels.
• Regime detection. A sustained rise in ATR relative to its own history indicates a change in market conditions, which many strategies use to reduce exposure.
• Stop-out probability estimation. Knowing the typical daily movement lets you estimate how likely a given stop is to be reached by noise alone.

Limitations and practical notes

• It is backward looking. ATR reports what movement has been, not what it will be. A sudden regime change is reflected only after it occurs.
• It is in price units. ATR of 2.00 means nothing without knowing the price. Divide by price for cross-instrument comparison.
• Smoothing choice matters slightly. Wilder smoothing responds more slowly than a simple average; results are similar but not identical.
• Short lookbacks are unstable. ATR(5) jumps around; ATR(20) is smoother. Most uses favour 14 to 20 periods.
• It ignores direction entirely. ATR rises in both crashes and rallies, which is correct behaviour but means it cannot inform which way to trade.
• In 24-hour markets, gaps are rarer, so true range approaches simple range and the gap-handling advantage diminishes.