5 Head and Shoulders Confirmation Rules

The five most useful head and shoulders “strategies” for a beginner to compare are five different confirmation rules: neckline wick crossing, candle close beyond the neckline, buffered close, break plus follow-through, and break plus retest. They are not five guaranteed ways to profit. They are five ways to define when a completed pattern is considered observable.

Each rule makes a trade-off:

earlier rules identify more events but include more temporary crossings;
later rules require more evidence but introduce delay;
retest rules exclude patterns that never return to the neckline;
volume or timeframe filters can change the sample rather than automatically improve it;
a measured move is an outcome hypothesis, not a promised target.

This page compares those rule sets and shows how to test them without changing the definition after seeing the result.

Educational note: This article explains chart-pattern research rules. It does not recommend a position, entry, stop, target, leverage level, or amount of capital to risk.

Use the Correct Head and Shoulders Page

ChartMini separates pattern definition from rule-set comparison.

Question	Owner
What are the left shoulder, head, right shoulder, neckline, and inverse version?	Head and Shoulders Pattern Guide
How do five possible confirmation rules differ?	This guide
How do common chart patterns compare at a glance?	Chart Patterns Cheat Sheet
How do I train general pattern recognition?	Pattern Recognition With Chart Replay
How does candle-by-candle replay work?	Bar Replay Beginner Guide
How should a neckline's line-versus-zone boundary and width be defined?	Support and Resistance Zone Width Rules

The indexed pattern guide owns the structural definition and identification workflow. This page assumes a candidate pattern has already been recorded under that definition.

Define the Pattern Before Comparing Strategies

A rule comparison is meaningless if every rule is applied to a different shape.

Before testing any of the five strategies, freeze these fields:

Field	Required definition
Pattern direction	Standard top or inverse bottom
Prior trend	Written uptrend or downtrend rule
Left shoulder	Pivot rule and timestamp
Head	Required distance beyond the shoulder
Right shoulder	Required relationship to the head
Neckline	Two anchor points and projection method
Neckline type	Horizontal, rising, or falling
Candle state	Closed bars only or an explicit live-bar rule
Data identity	Instrument, product, provider, session, timezone, price type, adjustment
Cutoff candle	Last candle available when the candidate is recorded
Failure before completion	Conditions that remove the candidate before a break is tested

Do not redefine the shoulders after later price appears. Do not rotate the neckline to capture a cleaner break. Do not move from wick-based anchors to close-based anchors only when one version gives a better outcome.

Standard and inverse patterns

The same five confirmation rules can be tested in both directions:

a standard head and shoulders uses a move below its neckline;
an inverse head and shoulders uses a move above its neckline.

Keep the two directions as separate fields in the dataset. Market behavior, short-selling constraints, gap behavior, and execution conditions may differ even when the geometric rules are mirrored.

The Five Confirmation Strategies at a Glance

Strategy	Confirmation event	Main advantage	Main limitation
1. Wick crossing	Any qualifying high/low moves beyond the neckline	Earliest event definition	Includes intrabar crossings that close back inside
2. Candle close	A completed candle closes beyond the neckline	Uses completed-bar evidence	Later than a wick event and still can recover
3. Buffered close	A close exceeds the neckline by a predefined distance	Filters marginal closes	Buffer choice can be arbitrary or hindsight-fitted
4. Follow-through	A break is followed by a predefined continuation condition	Requires evidence after the initial break	Adds delay and may exclude fast reversals or gaps
5. Retest	Price breaks, returns to a defined neckline area, then meets a hold/rejection rule	Separates break and return phases	Many patterns never retest; retests can fail

None is automatically “best.” Each strategy defines a different event and therefore produces a different dataset.

Strategy 1: Neckline Wick Crossing

Definition

For a standard top, the event occurs when a candle's low trades below the projected neckline. For an inverse bottom, the event occurs when a candle's high trades above it.

The rule must specify whether the neckline value is calculated at:

candle open time;
candle close time;
the exact timestamp of the intrabar extreme;
a fixed horizontal level;
a projected sloping line.

A sloping neckline creates an additional timing issue because its price changes across the candle. On OHLC data, the exact intrabar path is unknown. A candle can contain a high and low on both sides of the line without revealing which crossing occurred first.

What this rule measures

The wick rule answers:

Did the recorded price range extend beyond the neckline at any point during the candle?

It does not answer:

whether price remained beyond the neckline;
whether the candle closed beyond it;
whether a live order could have filled at the displayed price;
whether the crossing occurred before another intrabar event;
whether later candles followed through.

Useful outcome labels

After the wick event, classify later behavior separately:

closed beyond the neckline in the same candle;
closed back inside the pattern;
crossed again on the next candle;
recovered above or below the neckline within the review horizon;
continued without a close back inside;
insufficient data before the sample ended.

Main bias risk

The wick rule often looks attractive on completed charts because the earliest crossing is visually obvious after the move. During live observation, however, the same candle was incomplete and its final close was unknown. Mark the event as a wick event rather than retroactively calling it a confirmed break.

Strategy 2: Completed Candle Close Beyond the Neckline

Definition

For a standard top, a completed candle must close below the neckline. For an inverse bottom, it must close above.

The rule should define whether equality counts. Examples:

strict: close < neckline or close > neckline;
inclusive: close ≤ neckline or close ≥ neckline;
tick-aware: close must be at least one minimum price increment beyond;
percentage-aware: equality and rounding are handled under a stated precision rule.

What this rule changes

A close rule ignores wick-only crossings that recover before the candle completes. That can reduce the number of temporary events, but it does not guarantee continuation. Schwab's educational material explicitly notes that even a neckline breach is not a lock to continue in the same direction.

Timeframe dependency

A close is defined by the chosen timeframe. The same price sequence can produce:

a five-minute close beyond the neckline;
no hourly close beyond it;
a daily candle that closes back inside.

Do not call one timeframe the “real” confirmation after seeing the outcome. State the decision timeframe before testing.

Recommended record fields

break-candle timestamp;
neckline price at the close;
close-to-neckline distance;
candle range;
close location inside the candle range;
gap status;
volume and its data source, if used;
next-candle status;
review-horizon outcome.

These are observations, not automatic trading instructions.

Strategy 3: Buffered Close Beyond the Neckline

Definition

A buffered close requires the completed candle to exceed the neckline by a predefined distance.

Possible buffer units include:

ticks or minimum price increments;
pips;
absolute price units;
percentage of the neckline price;
fraction of recent median range;
fraction of a stated volatility measure.

The buffer must be chosen before the event is reviewed.

Why use a buffer?

A one-cent close below a neckline and a large close beyond it are technically both closes under a strict binary rule. A buffer creates a separate category for marginal crossings.

The buffer does not prove that a move is genuine. It only changes the required distance.

Buffer formula examples

A percentage buffer can be recorded as:

Break distance % = |Close - Neckline| / Neckline × 100

A range-normalized buffer can be recorded as:

Normalized break distance = |Close - Neckline| / Baseline range

The baseline range might be median true range over a predefined window or another stated measure. Avoid switching baselines between patterns.

Hindsight risk

Buffered rules are easy to overfit. An analyst can test many percentages, periods, timeframes, and volatility multipliers until one performs well on the same sample.

Use separate samples:

Development sample: choose and document the buffer.
Evaluation sample: apply the frozen rule without adjustment.

If the buffer changes, create a new rule version rather than overwriting the original.

Strategy 4: Break Plus Follow-Through

Definition

The event has two stages:

an initial neckline break under a stated wick, close, or buffered-close rule;
a follow-through condition within a stated number of later candles.

Possible follow-through definitions include:

next candle also closes beyond the neckline;
no close back inside for a fixed number of candles;
a new low after a standard top or new high after an inverse bottom;
cumulative movement beyond a stated distance;
a close beyond the break candle's extreme;
a sequence rule based on closes, not highs and lows.

Choose one. Combining several conditions after seeing the chart creates an unfalsifiable rule.

What this strategy measures

This strategy asks whether the initial break persisted under the chosen horizon. It is not the same as confirming the original pattern at the first break.

Record two timestamps:

initial break time;
follow-through confirmation time.

The delay between them matters. A rule that appears cleaner may identify the event after a large part of the later movement has already occurred.

Failure and no-result labels

initial break but no follow-through;
initial break followed by immediate recovery;
follow-through achieved within horizon;
follow-through achieved after horizon;
sample ended before classification;
data gap prevented classification.

Do not silently convert “achieved after horizon” into a success if the original rule required completion within the horizon.

Strategy 5: Break Plus Neckline Retest

Definition

A retest strategy requires:

a completed break under a predefined rule;
a later return to a predefined neckline area;
a predefined hold, rejection, or re-break condition.

Each stage needs its own definition.

Define the retest area

A sloping neckline is a moving reference. A retest can be defined as:

touching the projected line;
entering a fixed-width band around it;
closing inside the band;
crossing back through and closing beyond again;
reaching the original break point rather than the projected line.

Do not use the nearest visually convenient level after the return occurs.

Define the retest horizon

State how long the test remains eligible:

number of candles;
calendar time;
trading sessions;
until a specified structural event;
until the pattern is retired.

A return months later may not answer the same question as a return on the next candle.

Selection effect

Retest rules exclude patterns that never return to the neckline. That is not necessarily a defect, but it means the retest sample is not directly comparable with all-break samples.

Report:

percentage of completed breaks that produced an eligible retest;
time from break to retest;
maximum movement before retest;
retest-zone depth;
hold, cross, repeated-crossing, or no-result outcome.

A retest can hold, fail, overshoot, or never occur. It is not guaranteed confirmation.

How Volume Fits the Five Strategies

Volume is often taught as context for a neckline break, but it must be defined carefully.

Different markets provide different volume data

exchange-traded stocks and futures may provide venue or consolidated volume;
retail Forex commonly provides tick volume rather than centralized traded volume;
crypto volume is exchange-specific unless an aggregation method is used;
CFDs and synthetic products may use provider-specific data.

Do not compare these as if they were the same dataset.

Define the baseline

A phrase such as “high volume” is not reproducible. Record:

volume data source;
timeframe;
baseline window;
mean, median, percentile, or another comparison statistic;
whether the break candle is included in the baseline;
threshold chosen before evaluation.

Examples of testable labels include:

break volume above the prior 20-bar median;
break volume in the top quartile of a frozen reference window;
no volume filter;
volume unavailable or not comparable.

Volume changes the rule set. It does not guarantee continuation or reveal who caused the move.

How to Treat the Measured Move

A commonly taught head and shoulders measurement uses the vertical distance from the head to the neckline, projected from the break.

For a standard top:

Pattern height = Head price - Neckline price at the head

Projected level = Break reference - Pattern height

For an inverse bottom, the direction is reversed.

This is a measurement hypothesis, not a target guarantee. Several choices must be defined:

neckline price at the head, at the break, or another timestamp;
head high/low based on wick or close;
break reference based on wick, close, or buffered close;
treatment of a sloping neckline;
review horizon;
whether touching the projection or closing beyond it counts;
partial-distance thresholds;
intervening adverse movement.

A useful evaluation table can record:

Field	Example classification
25% of measured distance reached	Yes / No / No result
50% reached	Yes / No / No result
100% reached	Yes / No / No result
Time to each threshold	Candle count or elapsed time
Maximum favorable movement	Price and normalized distance
Maximum adverse movement	Price and normalized distance
Closed back through neckline	Timestamp or never within horizon

Do not report only cases that reached the full projection.

Compare the Five Rules on One Frozen Sample

A fair comparison uses the same candidate-pattern set.

Step 1: Build the candidate set

Use the indexed Head and Shoulders Pattern Guide to define the prior trend, shoulders, head, neckline, and pre-break invalidation.

Freeze candidates before evaluating later price. Include candidates that never complete, because completion rate is part of the evidence.

Step 2: Apply every eligible rule

For each candidate, record whether and when each rule fired:

Candidate	Wick	Close	Buffered close	Follow-through	Retest
Pattern A	Timestamp	Timestamp	No event	No event	No retest
Pattern B	Timestamp	Timestamp	Timestamp	Timestamp	Timestamp
Pattern C	No event	No event	No event	No event	Not eligible

Do not remove Pattern C merely because no rule completed.

Step 3: Use the same outcome horizon

Possible horizons include:

fixed number of candles;
fixed elapsed time;
next major structural event;
a maximum of both time and candles.

Changing the horizon for each rule can make the comparison misleading.

Step 4: Report event timing and coverage

For each rule, report:

number of eligible candidates;
number and percentage of triggered events;
median confirmation delay from first neckline contact;
percentage with no event;
percentage with retest eligibility;
percentage with insufficient data;
movement and recovery distributions;
data errors or excluded cases.

A later rule may appear to produce cleaner charts partly because it rejects more candidates and waits longer.

Define Pattern Failure Before Testing

Failure can occur at different stages.

Candidate failure before a neckline break

Examples:

right shoulder exceeds the head under the frozen rule;
prior-trend requirement is no longer met;
neckline anchors become invalid under the stated pivot method;
the pattern exceeds a maximum duration;
data correction changes the source pivots.

Break-event failure

Examples:

wick crosses but closes back inside;
close occurs beyond the line but returns inside within the failure horizon;
buffered close threshold is not met;
follow-through condition fails;
retest crosses through the defined band and meets the failure rule.

Outcome failure

An outcome rule must be separate from a pattern definition. For example, failing to reach a measured projection does not retroactively mean the geometric pattern never existed. It means the stated outcome was not observed within the horizon.

Common Beginner Errors

Calling the right shoulder a completed pattern

The right shoulder may create a candidate shape, but a completion rule has not fired yet. Label candidate, complete, and evaluated stages separately.

Mixing the five strategies

A common record says “wait for a close, preferably volume, maybe a retest, unless momentum is strong.” That cannot be tested consistently. Create separate rule versions.

Switching timeframes after a failure

A five-minute break that fails should not be reclassified as “not confirmed because the hourly candle stayed inside” unless the hourly-close rule was chosen in advance.

Treating a retest as inevitable

Some breaks do not return to the neckline. A retest-only rule measures a subset of completed breaks.

Optimizing the buffer on the same sample

Testing many percentages and keeping the best one creates data-snooping risk. Freeze the selected buffer before evaluating a new sample.

Assuming volume proves institutional participation

A volume bar records activity under the provider's methodology. It does not identify banks, funds, market makers, algorithms, or “smart money.”

Treating the measured move as an order instruction

The projection is one way to measure later movement. It does not determine a suitable position, stop, target, or risk level for an individual.

Ignoring incomplete patterns

If a candidate never crosses the neckline, it still belongs in the candidate dataset. Removing it inflates completion statistics.

A Head and Shoulders Rule-Comparison Worksheet

Use one record per candidate pattern.

Chart identity

instrument and product;
provider or venue;
price type;
session and timezone;
adjustment method;
timeframe;
lookback window;
candidate cutoff candle.

Pattern definition

standard or inverse;
prior-trend rule and result;
left-shoulder timestamp and price;
head timestamp and price;
right-shoulder timestamp and price;
neckline anchor timestamps and prices;
neckline slope and equation;
pre-break invalidation condition;
candidate duration.

Five strategy events

first wick crossing timestamp and distance;
first close crossing timestamp and distance;
buffered-close rule version and result;
follow-through rule version and result;
retest band, horizon, and result;
volume context and baseline;
data-quality notes.

Outcome review

review horizon;
closed back through neckline;
time to recovery or continuation threshold;
maximum favorable movement;
maximum adverse movement;
25%, 50%, and 100% measured-distance results;
no-result reason;
screenshot or replay reference;
rule changes prohibited after evaluation starts.

How to Practice the Five Strategies With ChartMini

ChartMini can hide future historical candles so the five rules are applied before the outcome is visible.

Open a historical chart in ChartMini.
Fix the instrument, data source, session, timeframe, and lookback.
Advance until a candidate head and shoulders or inverse pattern can be recorded.
Mark the shoulders, head, neckline anchors, and cutoff candle.
Write the five rule definitions before advancing.
Reveal one candle at a time.
Record the first event for each rule without moving the neckline.
Preserve candidates that never complete.
Continue to the same outcome horizon for every rule.
Compare timing, coverage, recoveries, retests, and measured-distance observations across many patterns.

ChartMini supports historical candle replay and visual annotation practice. It does not reproduce live bid/ask changes, order-book depth, queue position, slippage, commissions, financing, margin events, liquidity, or actual fills. A replay study cannot prove that a strategy will perform similarly in live trading.

Official Sources and Evidence Boundaries

Charles Schwab: Identifying Head-and-Shoulders Patterns in Stock Charts — explains that the pattern is easy to misread, identifies the neckline as the fourth component, and notes that a neckline breach does not guarantee continuation.
StockCharts ChartSchool: Head and Shoulders Top — provides the conventional prior-trend, component, neckline, volume, and measurement framework.
StockCharts ChartSchool: Head and Shoulders Bottom — documents the inverse version and its neckline/volume framework.
CFTC: Commodity Trading Systems Sold on the Internet — explains the inherent limitations of hypothetical and simulated trading results and warns that no trading system can guarantee profits.

These sources describe common technical-analysis conventions. They do not establish one universally superior confirmation rule. The five-rule framework on this page is a transparent comparison procedure, not an official performance standard.

Frequently Asked Questions

What is the safest head and shoulders confirmation strategy?

There is no universally safest confirmation rule. A wick break reacts earliest but includes more temporary crossings. A close, buffered close, follow-through rule, or retest rule requires progressively more evidence but can identify the event later or miss patterns that never retest. Compare the rules on the same historical sample before choosing one for further study.

Does a head and shoulders pattern require a neckline close?

Many technical-analysis frameworks treat a neckline break as the completion event, but the exact break rule must be stated. A wick through the neckline, a candle close beyond it, and a close beyond a buffer are different definitions and should not be mixed in one test.

Should beginners wait for a neckline retest?

A retest can provide a later observation point, but it is not guaranteed to occur or hold. Waiting for a retest changes the sample because patterns that continue without returning are excluded. Treat retest confirmation as one rule set to evaluate rather than a universal requirement.

How should volume be used with a head and shoulders pattern?

Volume can be recorded as context when the market and data source provide meaningful volume. Define the baseline, comparison window, and threshold before reviewing the outcome. Stock volume, futures volume, retail forex tick volume, and exchange-specific crypto volume are not equivalent datasets.

Does the measured move guarantee a head and shoulders target?

No. Projecting the head-to-neckline distance from the break is a commonly taught measurement hypothesis, not a guaranteed destination. Record whether and when the projection was reached, along with adverse movement, partial movement, and the review horizon.

Can ChartMini prove which head and shoulders strategy works best?

ChartMini can hide future historical candles so you can apply predefined rules and review later price action. It does not reproduce live order execution, bid-ask changes, slippage, liquidity, commissions, financing, or participant decisions. Replay results are practice evidence, not proof of future performance.

5 Head and Shoulders Confirmation Strategies Beginners Can Test