ATR-Based Position Sizing

Advanced Track. Complements fixed fractional risk in Risk Management 101 and fractional optimization in Kelly Criterion sizing. Strategy context from trend following, swing trading, and day trading. Hub: stock courses.

Same Dollar Risk, Different Volatility

A $1.00 stop on a quiet large-cap and a $1.00 stop on a high-beta name after an earnings gap are not equal risk experiences. The second name routinely moves $1.00 in minutes; the first may take a day. If you always buy the same share count, you are silent-leverage trading: volatile names dominate your P&L noise and your drawdowns. ATR-based position sizing makes volatility explicit — stops and share counts scale with how much the instrument actually moves — while your account risk budget (percent of equity) stays controlled.

This course defines Average True Range, shows how to place ATR-based stops, convert a fixed fractional (or Kelly-capped) dollar risk into shares, cut size by regime, and avoid the false precision of treating ATR as destiny. Free tools — especially the risk / position size calculator — remain the execution layer after ATR sets the stop distance.

1. What ATR Measures

Average True Range (ATR), popularized by Welles Wilder, is a moving average of true range — the greatest of: high−low; |high−prior close|; |low−prior close|. That definition captures gaps: overnight jumps count as range even if the session candle is small. ATR is not directional. It does not say long or short. It estimates how much noise and travel to expect, in price units, over the lookback (commonly 14 periods on the chart timeframe you trade).

Read ATR on the same timeframe as your stop decision. A daily ATR for swing stops; a 15-minute ATR for day-trade noise buffers. Mixing a daily ATR stop with a 1-minute entry without a multi-timeframe plan — see multiple timeframes — produces incoherent risk. Chart and structure literacy from chart reading and market structure still decide where the thesis dies; ATR helps decide how far is “beyond noise” and how many shares fit the budget.

2. The Core Formula: Risk Budget ÷ Stop Distance

ATR sizing is not magic. It is the same identity you already use:

Shares = Dollar risk budget ÷ Stop distance per share

ATR enters by setting stop distance (and sometimes by scaling the budget in high-vol regimes). Example swing framework:

• Dollar risk budget = 1% of equity (or half-Kelly capped — Kelly course)
• Stop distance = max(structure distance, k × ATR)
• Or: stop = entry − k×ATR for longs (with structure check)

Common multipliers: 1.5× to 3× ATR beyond a swing for swings; tighter fractions of ATR for day trades after structure is defined. There is no universal k — it is a playbook parameter you journal and test. Convert with the risk calculator; pre-plan exits with the SL/TP calculator.

Worked example A (swing). Equity $60,000. Risk 1% = $600. Daily ATR(14) = $2.40. Long entry $80.00. Structure swing low suggests $76.80 (distance $3.20). 2×ATR = $4.80. You choose stop at structure $76.80 because it is beyond 1×ATR noise but tighter than blind 2×ATR — still valid if structure is real. Distance $3.20 → shares = $600 / $3.20 = 187 shares. If you forced a $1 stop inside the ATR noise, you would own more shares and get shaken out by normal volatility — false precision.

Worked example B (high vol). Same budget $600. ATR jumps to $6.00 after an event. Structure stop is $9.00 away. Shares = $600 / $9 = 66 shares. Notional falls even though “conviction” may feel higher. That is the point: ATR sizing automatically de-levers when the market’s ruler lengthens.

3. Structure First, ATR Second

ATR alone can place a stop in mid-air — a level with no auction meaning. Professionals prefer: mark invalidation from structure, support/resistance, or pattern boundary (S/R deep dive, chart patterns), then verify the distance is at least a sensible ATR multiple so noise does not dominate. If structure is closer than ~1×ATR, either accept higher shakeout rate, wait for a better location, or skip. If structure is many ATRs away, size collapses — do not “need” the trade by ignoring the stop.

VWAP and session anchors from VWAP course often define intraday structure; ATR on the execution timeframe sets the buffer beyond that anchor. Breakout and trend systems from breakouts and trend following commonly trail with ATR once in profit.

4. ATR Trailing Stops

After entry, a trailing stop of the form “highest close since entry − k×ATR” (longs) adapts as volatility changes. In quiet trends, the trail tightens in price terms as ATR falls; in expansion, it loosens so you are not stopped by the trend’s own range. Partials still apply — bank process wins at structure targets, trail the rest — as in swing trading.

Do not trail so tightly that every pullback is a full exit if your edge is multi-day continuation. Do not trail so wide that a thesis break leaves a catastrophic R multiple. Journal k and timeframe. Use the P&L calculator when scaling out so partial math is intentional.

5. Regime Cuts: When ATR Explodes

ATR is a regime detector. When ATR doubles after earnings, FOMC, or a crash tape, automatic share reduction is a feature. Add discretionary cuts:

• Halve risk budget when ATR percentile is extreme vs its own history
• Reduce overnight risk further on high-ATR names (gap risk — gap trading)
• Stand down when spreads widen and ATR-implied stops are untradeable in thin books (microstructure)

Mean-reversion strategies from mean reversion often need different k than trend systems: too-wide ATR stops can turn a range fade into a trend hold against you. Regime identification first, ATR second.

6. Integrating ATR with Fixed Fractional and Kelly

Think in layers:

1. Policy risk %: fixed fractional or half-Kelly capped (Kelly, Kelly calculator)
2. Portfolio caps: total open and sector risk (correlation & portfolio risk)
3. Stop distance: structure + ATR logic
4. Shares: budget ÷ distance

ATR never increases your allowed percent of equity above policy. It only changes how that percent converts into shares. Momentum names with high ATR from RS/momentum should not receive the same share count as low-ATR index ETFs at the same risk percent — ATR sizing enforces that automatically.

7. Day Trading vs Swing: Timeframe of ATR

Day traders from day trading fundamentals typically use ATR on the execution timeframe (e.g. 5m or 15m) for noise buffers and OR-related stops, with smaller risk percents. Swing traders use daily ATR for structural trails and overnight risk. Using daily ATR to set a 5-minute stop is usually too wide for day-trade R:R; using 5-minute ATR to set a multi-day stop is usually too tight for swings. Match the clock to the hold — the same principle as multi-timeframe analysis in Course 20.

Session volatility patterns (open drive vs lunch) mean intraday ATR can change through the day. Some desks freeze ATR from a pre-market or prior-day value for consistency; others update. Pick one method and journal it. Broker and session mechanics: broker course, how markets work.

8. Targets, R:R, and ATR Multiples

Some systems set targets as +2×ATR while stops are −1×ATR for a theoretical 1:2. That can work in trends and fail in ranges. Prefer targets at structure (prior highs, measured moves, Fib extensions from Fibonacci) and use ATR mainly for stop and trail. If ATR-based stop makes R:R to the next structure worse than your minimum, skip the trade — do not invent a fantasy target mid-air.

Validate planned R with the SL/TP calculator and net outcomes with the P&L calculator. Track whether ATR-filtered trades improve expectancy with the win rate calculator and break-even calculator. Measure how extended a name already is with the percentage change calculator before applying wide ATR stops that allow huge notional by mistake (always compute shares from dollar risk, not the reverse).

9. Common ATR Sizing Mistakes

• ATR stop with no structure (noise-only invalidation)
• Wrong timeframe ATR for the hold period
• Using ATR to increase risk % “because vol is opportunity”
• Ignoring correlation: three high-ATR names still stack factor risk (portfolio risk)
• Changing k every trade to avoid a stop-out (curve-fitting in real time)
• Forgetting gaps: ATR includes gaps historically but one event can still exceed stop (gaps)
• Emotional override after a win streak (beginner mistakes)

Volume and RS context still apply: volume analysis, momentum & RS. Optional reading: DennTech blog. Venues: exchanges hub. Tool stack: all calculators.

10. Implementation Checklist

1. Choose ATR timeframe = stop timeframe.
2. Compute ATR(14) or your locked lookback.
3. Mark structural invalidation.
4. Set stop = structure beyond noise (compare to k×ATR).
5. Dollar risk = min(policy %, portfolio remaining budget) × equity.
6. Shares = dollar risk ÷ stop distance; floor to lot constraints.
7. Check liquidity and spread; cut if untradeable.
8. Define target/trail; log ATR, k, and regime tag.

11. Practice Drill (Fifteen Trades)

Pick one strategy (e.g. daily trend pullback). For fifteen trades, force: stop at least 1.5× daily ATR beyond the swing low/high used for structure, risk fixed 0.75% equity, no discretionary k changes. Compare stop-out rate and average R to your prior fixed-dollar-stop sample. If shakeouts fall without killing R:R, keep ATR. If R:R collapses because stops are too wide, improve location selection rather than secretly tightening stops after entry.

Foundations of ownership: What Is Stock Trading? Probability framing: Intro to TA. Moving averages as dynamic structure often pair with ATR trails: moving averages. Support at session levels: S/R, VWAP.

12. Narrative: Two Names, One Risk Percent

A trader risks 1% on SPY and 1% on a high-ATR biotech the same morning with equal share counts “for simplicity.” The biotech’s ATR is 8× SPY’s on a percentage basis. By noon the biotech alone has consumed more emotional and dollar volatility than the entire SPY plan. ATR sizing would have cut biotech shares by roughly that volatility ratio, keeping both legs near the intended risk contribution. The lesson is not “never trade biotech”; it is never let share-count laziness override volatility-aware budgeting.

Second narrative: after an FOMC spike, ATR doubles on a swing candidate. The trader keeps yesterday’s share count and uses a tight discretionary stop inside the new noise band. Three stop-outs later, they blame the market. Correct process: recompute shares from the new ATR-aware stop distance, or pass until location offers a tighter structure with acceptable R:R. Volatility expansion is information, not an insult.

13. Governance: Lock Parameters, Review Quarterly

Write into your plan: ATR length, default k by strategy, when to freeze ATR (e.g. use prior day’s daily ATR for the entire session), and maximum risk % regardless of ATR. Review quarterly alongside Kelly inputs if you use them. Changing ATR length every week to fit recent trades is curve-fitting. Advanced operators change parameters slowly, with sample sizes that justify the change — the same statistical humility required for Kelly estimation error in Course 31.

Portfolio governance from Course 37 still applies: sum ATR-sized legs by sector. Three “correctly” ATR-sized semis can still be one crowded bet. ATR solves volatility normalization across names; it does not solve factor concentration.

Key Takeaways

Tools for This Course

• Risk & Position Size Calculator — convert ATR-aware stop distance into share count from a fixed dollar budget.
• Kelly Criterion Calculator — optional ceiling for risk % before ATR conversion.
• SL/TP + P&L — verify R:R after ATR stop is set.
• Stock Courses Hub — Track 4: Kelly (31), ATR (32), psychology, plans, flow, correlation, and more.