An inside candle fits within the preceding candle’s range. That geometry identifies a candidate; a trade still needs a defined breakout side, trigger, stop and expiry rule. This guide shows how to record those choices before judging the result, so a later winner does not decide which rules you pretend to have used.

What an inside bar is and why it matters in price action

An inside bar compares a candle’s full High–Low range with the preceding mother bar. In this article’s strict definition, H₂<H₁ and L₂>L₁; equality excludes the pattern. Other conventions permit touching one boundary and need a separate label. Body colors and relative body lengths do not determine this test. A narrower observed range alone does not forecast the size of the next move.

An inside bar contains the second candle's entire high-to-low range within the first candle's range. A harami instead compares the real bodies; the second candle's shadows can extend outside the first candle's range. Consequently, a harami is not necessarily an inside bar, and an inside bar is not necessarily a harami. Check both definitions separately.

Criteria for a classic inside bar
Mother barthe preceding candle selected as the range reference
Inside candleboth the low and the high sit strictly inside the mother bar's range
Body coloursirrelevant — may match or oppose the mother bar
Range ratiorecord (inside High−Low)/(mother High−Low); do not infer future volatility
Contexta separately defined filter to test, not a required geometric property
Visual silhouettea small candle tucked beneath the shoulders of a wider preceding bar

The mother bar and the mechanics of volatility compression

Use the observed ranges before inferring participant motives. Hypothetical EUR/USD Bid candles on H4 have mother O/H/L/C=1.1000/1.1040/1.0960/1.1020 and inside O/H/L/C=1.1020/1.1030/1.0990/1.1000. Their ranges are 80 and 40 pips, with pip size 0.0001, while both bodies are 20 pips. The full-range ratio is 50%, despite equal body lengths. Recognize the candidate only after the second candle closes.

Inside containment differs from NR4. Take three earlier ranges 10,30,20 units, with immediate predecessor High/Low=100/80. A next range 98/83 is inside but is not NR4: width 15 exceeds 10. A next range 111/103 has width 8, below all three earlier widths, but is not inside. These are hypothetical strict tests; record both Boolean results rather than renaming one filter as the other.

A release calendar can be a separate filter, but candle geometry does not identify why participants traded or waited. In the model below, entry activation is checked on Ask while the chart is Bid. Ask 1.1043 with a 2-pip spread means Bid 1.1041, one pip above mother High 1.1040. With a 5-pip spread it means Bid 1.1038, two pips below. A fixed Ask trigger therefore need not coincide with a Bid-chart breakout.

Why location on the chart matters more than the pattern itself

Context can be recorded as explicit inputs: a trend rule, distance to a chosen zone or upcoming announcement. Do not infer actual defending orders or a durable edge from a candle’s location. Fix each filter before testing, retain rejected cases and compare full outcomes. A trend-following and a counter-trend variant are different hypotheses.

Context fields to record — not performance rankings
Middle of a consolidation, no support or resistance nearbystate the range-location rule and whether this candidate passes
Aligned with the H4 trenddefine the H4 trend filter using information already available
Aligned with the daily trendrecord whether the chosen daily-trend condition passes
At a meaningful support or resistance zonerecord the zone’s construction and distance; do not infer who watches it
Trend, zone and compression at oncerecord the joint conditions and count qualifying and rejected cases
Timeframesdifferent periods produce different samples; test costs and rules without assuming a universal best interval

A reported win rate needs its sample, feed, interval, entry, exit and costs. Range containment by itself is not a completed trade. Compare the fixed rule on a separate period without adjusting rejected cases after seeing their outcomes. The source definitions here support measurement and order mechanics; they do not provide a new FX success-rate study.

The breakout strategy step by step

Choose a directional policy explicitly. This hypothetical research model uses only a buy stop, submitted after the strict inside candle closes and only while current Ask remains below the trigger. Adding an opposite sell stop creates another variant requiring stated cancellation and rapid-double-trigger handling. Fuller’s description also permits different stop placements; it does not make a far-side stop mandatory for every method.

  1. Pattern identification. Wait for the chosen candle period to close, then compare its entire high-to-low range with the preceding candle. Daily session boundaries depend on the broker. If a feed uses a 17:00 New York boundary, convert it for the actual date and compare it with the server timestamps; do not assume every broker uses that boundary.
  2. Context analysis. Check whether the trade direction matches the chosen trend filter, whether the pattern is near a meaningful support or resistance zone, and whether the entry avoids the middle of an unstructured consolidation. Define the required conditions before testing; an undesirable location must not count as a positive point.
  3. Place the chosen order. The model requests a buy stop at Ask 1.1043: mother Bid High 1.1040 plus an assigned 3-pip buffer. The buffer is not a guarantee of a visible Bid breakout when spread changes.
  4. Stop loss and position size. Define the invalidation level and buffer before entry, then calculate the planned monetary risk. A EUR10,000 account with a one-percent risk budget and a 50-pip stop allows EUR2 per pip. For a standard 100,000-unit EUR/USD lot, one pip is USD10, so convert it to the account currency: at an illustrative EUR/USD rate of 1.10, EUR2 per pip corresponds to 0.22 lot. Round down to the permitted volume step and account for fees and possible slippage; an ordinary stop does not guarantee a one-percent maximum loss.
  5. Order expiry. Cancel an unfilled entry after two subsequent candle periods, without renewing it for nested bars. This is the model’s chosen expiry, not a universal loss of pattern information. Two H4 data periods need not equal eight elapsed hours across a quotation break. A filled position uses a separate four-bar time exit, counting its fill bar as the first.

Take profits and position management after entry

Define target and exit policies before the test. For this model, a protective stop is requested on Bid 1.0958, two pips below mother Low, and the target on Bid 1.1123, entry trigger plus the mother’s 80-pip range. These are assigned levels, not provider guarantees. Do not recalculate them after worse entry unless that is a separately stated policy.

  • A fixed zone. Record its construction and the resulting reward-to-risk arithmetic. Geometry and distance alone do not establish positive expected return.
  • A range multiple. Specify the multiplier and starting point. A chosen two- or three-range target is a hypothesis, not a universal Fuller rule or guarantee of reach.
  • An ATR-based target. Specify the True Range formula, smoothing, period and multiplier. ATR is not simply the average High–Low range; validate this alternative separately.
  • A moving stop. Specify when it updates and how its level is calculated. Moving a stop to the entry price does not guarantee a net break-even result: spread, other charges and worse execution still matter.

Partial exits and trailing exits are distinct policies whose net outcomes depend on the entire path, fills and charges. Do not assume splitting a position always lowers monthly variability. Compare the same qualifying sample under each frozen policy instead of choosing the exit after seeing which path won.

Hypothetical EUR/USD execution outcomes

Alternative fills — EUR 10,000, USD account, total commission USD 2
Trigger / stop / targetAsk 1.1043 / Bid 1.0958 / Bid 1.1123
Target fillAsk entry 1.1043 → Bid 1.1123: +USD 78 net
Requested stop fillAsk entry 1.1043 → Bid 1.0958: −USD 87 net
Worse stop fillAsk entry 1.1043 → Bid 1.0953: −USD 92 net
Gap entry, then targetAsk entry 1.1048 → Bid 1.1123: +USD 73 net
Time exitAsk entry 1.1043 → Bid 1.1030: −USD 15 net

All rows are separate hypothetical endings, calculated as 10,000×(closing Bid−entry Ask)−USD 2. The Bid/Ask fills already account for spread; financing and other fees are excluded. Planned trigger-to-stop distance is 85 pips, not the mother’s 80. A gap entry at 1.1048 leaves 75 pips to the unchanged target. An ordinary stop is not a maximum-loss guarantee. If a candle contains entry, stop and target levels, OHLC alone does not establish which was touched first; use more detailed data or a stated conservative assumption.

Inside bars compare complete candle ranges; harami patterns compare real bodies. Neither definition settles the direction of the next move.

Three labels do not mean three exposures

A fictitious mother range 90–110 contains ranges 94–108,96–106 and 98–104. Each is also narrower than the previous range.

Candle count versus order countThe mother range 90–110 contains three nested ranges 94–108,96–106,98–104. All share one mother. One 0.10-lot position with a 20-pip move gives a 20 USD price change; three identical orders give 60 USD.ONE MOTHER, THREE SMALLER CANDLESMotherIB1IB2IB33 geometry labels ≠3 orders.One 0.10 lot: model 20 USD.Three 0.10 lots: model 60 USD.

Hypothetical price diagram. Separate exposure calculation: EUR/USD,10 USD/pip per lot,20 pips from each fill to a shared exit, without fees or slippage. Nial Fuller — Inside Bar confirms that several nested inside bars may share one mother bar.

Shared boundaries do not create separate risk units.

A rule allowing one pending order per mother does not become a three-order rule when IB2 and IB3 appear. Store both mother ID and order ID; another geometry label may refer to the same pending setup. If a separate variant sends three 0.10-lot orders and all fill, total volume is 0.30 lot. At the specified 20 pips, price change is 60 rather than 20 USD. This is neither maximum loss nor a fill forecast; it shows the effect of counting signals as independent positions. The original expiry time remains unchanged.

Five most common inside-bar trading mistakes

Review errors against the specified rule. The same geometry can feed several strategies, so a deviation from this model should not be presented as a universally invalid way to trade the pattern. Missing trigger data, inconsistent expiry and miscounted exposure are observable faults.

  • Changing selection after seeing outcomes. Record filters beforehand, retain rejected cases and assess the complete rule’s costs and results.
  • Changing the range reference. If the rule uses the original mother, do not silently move thresholds to nested candles. An inside-candle breakout is a separately defined variant, not necessarily meaningless.
  • Choosing a stop only to enlarge quantity. State the exit policy and calculate costs and adverse fills. Fuller describes both opposite-end and some midpoint placements, so “never inside” is not a universal definition.
  • Mixing directional variants. A trend filter and a counter-trend rule need separate records and tests. Do not call either superior from candle anatomy alone.
  • Claiming an interval guarantees quality. Measure actual candidate counts, costs and performance for the chosen period and feed. Shorter intervals are not mathematically devoid of information.

Practice plan — fixed rules and a separate test

Begin with repeatable identification and an explicit order record. A practice timetable is an organizational choice, not evidence that a few months makes a trader ready. Separate rule development from validation on data not used to choose the filters.

  1. First month. Choose one pair and daily session boundary. Review six months, identify every inside bar under the same strict definition and record the actual count without assuming a target frequency.
  2. Second month. Categorise each pattern by location: middle of a directionless consolidation, aligned with the H4 trend, aligned with the daily trend, at a meaningful support or resistance zone, and finally several of those conditions at once. Then check how price behaved in the forty-eight hours after each one.
  3. Evaluate the development sample. Measure activation, no-trade cases and net results for each recorded variant. Similar outcomes may reflect no measurable distinction; do not automatically loosen a filter until a preferred category wins.
  4. Test without retuning. Apply chosen rules to a separate period or forward demo observations, retaining every qualifying case. Neither twenty trades nor sixty days is a universal statistical-validity threshold.
  5. Before live trading. Test the complete rules and costs on historical and forward demo data. If you later use a small live position, define loss limits, execution checks and review triggers in advance. Suspend trading when those limits or operational checks fail; do not require one hundred live trades regardless of losses or changing conditions.

What to do when you close this page

Geometry is reproducible only after the equality convention, price series and period boundary are specified. Other choices can also be defined mechanically rather than remaining inherently discretionary. Use the following record to keep the pattern, order and result separate.

  1. Find ten inside candles on the daily chart. Open EUR/USD on the daily timeframe, scroll back six months and mark every candle whose high and low both sit inside the previous one. Next to each, write down two numbers: the mother bar's high and low. Those are your future entry and invalidation levels — and you will see how far they sit from the edges of the inside candle itself.
  2. Record selection results. Apply the predeclared filters and count actual acceptances and rejections. A handful remaining is not a required outcome or proof of an edge.
  3. Calculate planned exposure. Use an explicitly chosen account-currency budget, actual contract/pip value, stop assumptions, charges and volume step. Round down; if the minimum quantity exceeds the budget, record no feasible order under those assumptions. This is not a recommendation of a one-percent budget or proof an interval is globally unsuitable.

Related reading: pin bar and reversal-pattern trading — a complementary single-candle pattern with different mechanics; how to trade the breakout strategy for broader context on breakouts; how to draw support and resistance — without that skill, judging the context of an inside candle comes down to guesswork.