Evening Star Candlestick: Rules and Confirmation

An evening star candlestick is a three-candle bearish reversal candidate considered after a defined price advance. A practical reference version has a relatively large bullish first candle, a smaller middle candle, and a bearish third candle that closes into the first candle's body. The pattern records three completed OHLC intervals. It does not reveal who traded, why they traded, or whether price must continue lower.

The most important distinction is between pattern completion and later confirmation. The third candle can complete an evening star definition, while a separate rule may require a later close below the pattern low, follow-through, or a failed retest. Mixing those events produces inflated or inconsistent statistics.

Educational note: Candlestick analysis is technical analysis, not a guarantee of direction or profit. Historical and simulated observations can differ materially from live trading because of spread, slippage, liquidity, order timing, fees, market impact, and discretionary decisions.

Key Takeaways

An evening star is a three-candle sequence, not a single star-shaped candle.
The prior advance, first-candle size, middle-body size, gap rule, third-candle penetration, and expiry window all need explicit definitions.
Traditional gap rules may not transfer cleanly from exchange-traded daily charts to Forex, crypto, futures, or intraday charts.
The third candle can complete the pattern, while later candles may provide additional confirmation or failure evidence.
Five confirmation versions can be compared without declaring one universally best.
Body color, volume, resistance, indicators, and higher-timeframe context are variables to record, not proof of reversal.
Failed, expired, ambiguous, and unresolved candidates must remain in the sample.

Which Candlestick Guide Should You Use?

ChartMini separates candlestick literacy, general pattern catalogs, and individual-pattern testing.

Question	Primary guide
How are standard OHLC candles constructed and read in sequence?	Complete candlestick reading workflow
What are the main named candlestick patterns?	Candlestick patterns guide
How can general pattern filters and confirmation variables be compared?	Candlestick setup and confirmation checklist
How is a single-candle Shooting Star defined and invalidated?	Shooting Star rules
How should an Evening Star's three candles, gaps, penetration, confirmation, and failure be recorded?	This guide
How should price interaction with a resistance level be classified?	Key-level decision framework
How can pattern recognition be practiced without seeing future candles?	Chart replay pattern-recognition guide

This page owns the three-candle Evening Star definition, five confirmation versions, failure labels, and replay record. It does not replace the broad candlestick catalog or the single-candle Shooting Star page.

What the Three Candles Actually Record

A standard candle records four values for one interval:

Open, High, Low, Close

CME Group explains that candlestick charts visualize the same open, high, low, and close data as bar charts. The chart data are objective records for the selected feed and interval; pattern interpretation begins when rules are applied to those values.

For candle i:

Rangeᵢ = Highᵢ - Lowᵢ
Bodyᵢ = |Closeᵢ - Openᵢ|
Body shareᵢ = Bodyᵢ ÷ Rangeᵢ

Use a zero-range guard when Highᵢ = Lowᵢ.

For a bullish first candle:

Close₁ > Open₁
First body height = Close₁ - Open₁
First body midpoint = (Open₁ + Close₁) ÷ 2

For a bearish third candle:

Close₃ < Open₃

These equations describe geometry. Labels such as exhaustion, hesitation, trapped buyers, institutional distribution, or aggressive selling are interpretations that require separate evidence and should not be presented as facts derived from three OHLC bars.

A Reference Evening Star Definition

StockCharts describes the evening star as a bearish reversal pattern containing a long bullish candle, a small second candle, and a long bearish third candle that provides bearish confirmation. TradingView's built-in Evening Star definition also uses an uptrend, a small middle candle, and a bearish third close below the midpoint of the first body.

A reproducible reference version can be written as follows:

Candle 1: bullish expansion

Close₁ > Open₁
Body₁ ≥ first_body_threshold

The threshold must be defined. Options include:

body share of candle range;
body relative to the previous N candle bodies;
body relative to ATR or another volatility estimate;
a percentile within a development sample.

Do not use words such as large or strong without a measurement rule.

Candle 2: the star

Body₂ ≤ small_body_threshold

Possible denominators include:

Body₂ ÷ Body₁
Body₂ ÷ Range₂
Body₂ ÷ median body over the previous N candles

The second candle may be bullish, bearish, or near-zero body, depending on the version. A near-zero body should be recorded separately because it may qualify as an Evening Doji Star under some definitions.

Candle 3: bearish penetration

Close₃ < Open₃
Penetration = (Close₁ - Close₃) ÷ (Close₁ - Open₁)

For a bullish first candle:

Penetration = 0 means candle 3 closed at candle 1's close.
Penetration = 0.5 means candle 3 closed at the midpoint of candle 1's body.
Penetration = 1 means candle 3 closed at candle 1's open.
Penetration > 1 means candle 3 closed below candle 1's open.

A common reference threshold is Penetration ≥ 0.5. A test may choose a different value, but the threshold should be selected before results are reviewed.

Verify the Chart Before Classifying the Pattern

The same symbol can produce different candles under different chart settings. Record the chart identity before measuring the sequence.

Field	What to record	Why it matters
Instrument	Exact stock, ETF, futures contract, Forex pair, index, or crypto product	Similar symbols can represent different products
Venue or provider	Exchange, broker, consolidated feed, or data vendor	Highs, lows, opens, and closes can differ
Price type	Last, bid, ask, midpoint, mark, index, settlement, or adjusted price	The selected series changes candle geometry
Session	Regular, extended, overnight, continuous, or custom	Session boundaries determine gaps and daily opens
Timezone	Exchange, local, UTC, or platform setting	Daily candles may start and end at different times
Timeframe	Time interval or other bar construction	A daily pattern and a one-minute pattern compress different activity
Adjustment	Raw, split-adjusted, dividend-adjusted, or continuous-contract method	Corporate actions and rolls can create artificial gaps
Bar state	Closed or realtime	The newest candle can change until confirmed

Do not classify an evening star until candle 3 is closed under the selected chart definition.

Define the Prior Advance

An evening star is traditionally considered after an uptrend or advance. “Price was going up” is too vague for testing.

Choose one rule and keep it fixed.

Pivot-sequence version

Require a defined series of higher highs and higher lows before candle 1.

Record:

pivot algorithm;
equality tolerance;
confirmation delay;
local versus major structure;
the last confirmed swing used.

For detailed swing and break terminology, use the market structure verification guide.

Net-change version

Prior return = (Close before candle 1 - Close N bars earlier) ÷ Close N bars earlier

Require the prior return to exceed a predefined threshold.

Range-break version

Require price to close above a previously frozen range high before the three-candle sequence begins.

Slope version

Require the slope of a predefined average, regression line, or closing-price series to be positive over a fixed lookback.

A moving average may be used to classify trend, but its period and comparison rule are part of the test. There is no universal reason that a 20 EMA, 50 SMA, or 200 SMA must define every evening star context.

Does an Evening Star Require a Gap?

Gap rules are one of the largest sources of inconsistent Evening Star counts.

Full-range gap

A strict upward full-range gap between candle 1 and candle 2 requires:

Low₂ > High₁

This is relatively uncommon in many continuously traded markets and intraday datasets.

Body gap

A less strict body-gap rule can be written as:

min(Open₂, Close₂) > max(Open₁, Close₁)

The bodies do not overlap, even if the wicks do.

Higher-body version

A looser version may require the second body to sit higher than the first body without requiring a complete gap.

No-gap version

A continuous-market version may omit the gap requirement and rely on:

a defined prior advance;
a relatively large first bullish body;
a small second body;
a bearish third candle meeting the penetration rule.

These are different pattern versions. Do not compare results from a strict stock-market gap definition with a no-gap Forex definition as though they describe the same candidate set.

Handle the Middle Candle Explicitly

The middle candle is frequently described as indecision. The OHLC data only establish that its open and close are close under a chosen threshold.

Record:

bullish, bearish, or equal-body color;
body size and body share;
upper and lower wick lengths;
body gap and full-range gap status;
whether its high is the pattern high;
whether it qualifies as a Doji under a separate epsilon rule.

A useful near-zero-body rule is:

Body₂ ÷ Range₂ ≤ doji_epsilon

The epsilon must be stated. Avoid dividing by zero when the candle range is zero.

An Evening Doji Star can be analyzed as a separate subgroup. It should not be silently combined with every small-bodied middle candle if the research question is about doji-specific behavior.

Distinguish Similar Bearish Patterns

Pattern	Number of candles	Core geometry	Main distinction
Evening Star	3	Bullish first, small middle, bearish third penetrating first body	Three-candle sequence with an explicit middle star
Evening Doji Star	3	Evening Star with near-zero middle body	Requires a doji threshold
Shooting Star	1	Small body near range low with long upper wick after an advance	Single-candle geometry; later confirmation is separate
Bearish Engulfing	2	Bearish second body engulfs bullish first body	No middle star candle
Dark Cloud Cover	2	Bearish second candle closes into bullish first body	Two-candle penetration pattern
Bearish Abandoned Baby	3	Doji isolated by gaps from candles 1 and 3	Much stricter gap structure
Morning Star	3	Bearish first, small middle, bullish third	Bullish mirror structure after a decline

The Shooting Star guide owns the single-candle upper-wick pattern. This page owns the three-candle Evening Star sequence.

Pattern Completion Is Not the Same as Additional Confirmation

The third candle often serves two roles in casual explanations:

it completes the three-candle pattern;
it is called confirmation of a possible reversal.

For testing, keep those roles separate.

Candidate identified after candles 1 and 2
Pattern completed after candle 3 meets the frozen definition
Additional confirmation evaluated only after completion
Outcome measured after the confirmation decision

This prevents future candles from changing whether the original three-candle sequence qualified.

Five Evening Star Confirmation Rules to Test

The following are rule versions, not recommendations. Compare them on the same frozen candidate set.

Rule 1: Midpoint Penetration Completion

The baseline version completes when candle 3 is bearish and closes below the midpoint of candle 1's body.

Close₃ < Open₃
Close₃ ≤ (Open₁ + Close₁) ÷ 2

Record completion at candle 3's close.

Trade-off:

earlier classification;
larger candidate set;
no evidence yet that price can move below the three-candle low.

This rule is close to several common educational definitions, including TradingView's built-in description.

Rule 2: Full-Body Penetration

A stricter completion version requires candle 3 to close at or below candle 1's open.

Close₃ ≤ Open₁

Trade-off:

deeper bearish penetration;
fewer candidates;
later or less favorable hypothetical decision prices;
still no guarantee of further decline.

Do not retroactively switch from midpoint penetration to full-body penetration only when the midpoint version fails.

Rule 3: Next Close Below the Pattern Low

The three-candle pattern completes under the chosen base definition, but additional confirmation requires a later completed candle to close below:

Pattern low = min(Low₁, Low₂, Low₃)

Trade-off:

requires price to move below the entire three-candle range;
confirmation occurs later;
some completed patterns never confirm;
gaps through the reference may create execution ambiguity.

Specify whether confirmation must occur on the next candle or within a fixed number of bars.

Rule 4: Break Plus Follow-Through

Require both:

a completed close below the pattern low;
a later bar that does not immediately close back above the reference and meets a predefined follow-through condition.

Possible follow-through definitions include:

Close remains below pattern low for K completed bars
Lowest close extends by X volatility units
No close returns above a predefined recovery reference

Trade-off:

reduces immediate failed-break confirmations;
delays the decision;
increases the number of expired candidates;
introduces another parameter that must be validated out of sample.

Rule 5: Failed Retest After the Break

After a close below the pattern low, wait for price to revisit a predefined reference and then close back below it.

Possible retest references include:

pattern low;
candle 3 body midpoint;
candle 3 open;
candle 1 body midpoint;
a separately frozen resistance zone.

A failed retest needs a precise rule. Merely touching a level and later declining is not enough if the retest and rejection definitions were not specified beforehand.

Trade-off:

may provide a clearer invalidation reference;
many valid declines do not retest;
a retest can occur after a large move;
hindsight can make the chosen reference look artificially accurate.

Location Is a Variable, Not Proof

StockCharts notes that bearish reversal candles may help confirm resistance after an advance. Record location categories before measuring outcomes.

Possible categories:

no nearby reference;
prior swing high;
horizontal resistance line;
predefined resistance zone;
upper range boundary;
calculated reference such as a moving average or Fibonacci level;
multiple overlapping references.

The presence of resistance does not prove that sellers are defending it. Use the key-level decision framework to distinguish approach, touch, rejection, breakout, retest, and failure without redrawing the level after the result.

Body Color and Candle Size Need Testing, Not Stories

The middle candle's color is often treated as unimportant, while the first and third candle directions are usually part of the pattern definition. That convention can be tested rather than assumed.

Record candidate subgroups:

bullish middle body;
bearish middle body;
near-zero middle body;
candle 3 body larger than candle 1 body;
candle 3 body smaller than candle 1 body;
candle 3 close above, at, or below the first-body midpoint;
candle 3 close above, at, or below the first open.

Compare subgroup results only when sample sizes are sufficient and the grouping rules were not selected after reviewing performance.

Volume Is an Optional Data Field

Volume can add context, but volume definitions differ across markets.

Record:

source and venue;
share, contract, tick, or proxy volume;
regular-session versus extended-session treatment;
comparison baseline;
whether the volume condition is part of classification, confirmation, or later analysis.

Example normalized fields:

Volume ratio₁ = Volume₁ ÷ median volume over previous N bars
Volume ratio₃ = Volume₃ ÷ median volume over previous N bars

Do not claim that higher third-candle volume proves institutional selling. It shows more recorded activity under the selected volume source.

Indicators and Higher Timeframes Are Separate Features

RSI, moving averages, channels, trendlines, Fibonacci levels, and sector indexes can be recorded as contextual variables. They should not silently redefine the Evening Star itself.

For each feature, state:

formula and settings;
source timeframe;
alignment method across timeframes;
threshold;
missing-data treatment;
whether the feature was selected before testing.

There is no universal evidence that RSI 70, a 20 EMA, a 61.8% retracement, or one specific higher timeframe transforms every evening star into a high-probability setup.

Define Failure, Invalidation, and Expiry

These labels should be written before the replay begins.

Disqualified before completion

Examples:

candle 1 does not meet the bullish-body rule;
candle 2 is too large;
the required gap is absent;
candle 3 is not bearish;
candle 3 does not meet the penetration threshold;
the prior-advance condition is absent.

A disqualified sequence is not a completed Evening Star failure. It never met the definition.

Completed but unconfirmed

The three-candle sequence meets the base definition, but the selected additional confirmation rule does not occur.

Confirmed but failed

The additional confirmation occurs, then a predefined failure event follows.

Possible failure references:

Close above pattern high
Close above candle 3 open
Close above candle 1 body midpoint
Return above the broken pattern low after a downside break
Adverse move threshold reached before the favorable threshold

Expired

No confirmation or failure occurs within the maximum holding or observation window.

Ambiguous

The available bar data do not establish whether a favorable or adverse threshold occurred first inside the same candle.

Do not delete ambiguous observations. Change the data resolution or keep the result unresolved.

OHLC Bars Do Not Reveal Every Intrabar Event

Suppose a later candle has:

High above the failure threshold
Low below the favorable threshold

A standard OHLC candle does not reveal which threshold was reached first. Declaring the trade a win or loss based on whichever result is preferred introduces bias.

Possible controls:

use lower-timeframe data that preserve the event order;
classify the observation as ambiguous;
apply a conservative tie rule fixed before testing;
exclude all same-bar dual-threshold events under a documented rule.

The same issue applies to gaps, stop orders, limit orders, and assumed fills.

Separate Classification, Confirmation, and Outcome

Use distinct fields.

Pattern classification

candidate;
completed;
disqualified;
ambiguous.

Confirmation status

not evaluated;
confirmed under Rule 1, 2, 3, 4, or 5;
expired without confirmation;
confirmation ambiguous.

Outcome status

downside threshold first;
upside failure threshold first;
neither within horizon;
both in same bar;
missing data;
unresolved.

This structure prevents a pattern from being renamed only because the later price path was unfavorable.

Define Outcome Measures Before Testing

A universal win rate does not exist without a complete specification.

Possible measurements include:

Fixed-horizon return

Return_H = (Close at H - reference price) ÷ reference price

For a bearish hypothesis, lower returns may be classified as favorable under the chosen threshold.

Maximum favorable excursion

The largest downside movement from the reference price during the evaluation horizon.

Maximum adverse excursion

The largest upside movement from the reference price during the same horizon.

Threshold-first outcome

Which predefined favorable or adverse level occurred first?

Time to event

How many completed bars elapsed before confirmation, failure, favorable outcome, or expiry?

Expectancy after costs

If a simulated entry and exit rule is used:

Expectancy = Win rate × average net win - Loss rate × average net loss

Use net outcomes after the stated spread, slippage, commissions, fees, and rejected or unfilled orders. A high classification accuracy does not prove positive expectancy.

Why There Is No Universal Evening Star Win Rate

A reported Evening Star result changes when any of the following changes:

market and instrument;
date range;
survivor universe;
timeframe and session;
feed and price type;
adjustment method;
prior-advance rule;
first-body threshold;
second-body threshold;
gap requirement;
third-candle penetration rule;
confirmation rule;
failure rule;
holding horizon;
transaction costs;
treatment of overlapping and ambiguous signals.

A percentage without those fields is not portable evidence. It should not be converted into “the pattern works” or “the pattern fails.”

Worked Classification Example

Assume three completed daily candles:

Candle 1: O=100, H=112, L=99, C=110
Candle 2: O=111, H=114, L=109, C=112
Candle 3: O=111, H=112, L=103, C=104

Calculations:

Body₁ = 10
First body midpoint = 105
Body₂ = 1
Body₂ ÷ Body₁ = 0.10
Body gap condition = min(111,112) > max(100,110) = true
Penetration = (110 - 104) ÷ 10 = 0.60

Under a reference version requiring:

bullish candle 1;
Body₂ ÷ Body₁ ≤ 0.30;
a body gap;
bearish candle 3;
penetration of at least 0.50;

this sequence completes an Evening Star at candle 3's close.

It is not yet confirmed under a rule requiring a later close below the pattern low:

Pattern low = min(99,109,103) = 99

A later close below 99 would be a separate confirmation event. If price closes above the predefined pattern high first, the sequence may be classified as failed under that version.

The numbers are illustrative. They do not imply a trade recommendation or expected return.

A Repeatable Evening Star Checklist

Before revealing later candles, record:

Exact instrument and product.
Venue or provider.
Price type.
Session and timezone.
Timeframe.
Adjustment method.
Closed-bar status.
Prior-advance rule.
Candle 1 body threshold.
Candle 2 small-body threshold.
Doji epsilon.
Gap version.
Candle 3 bearish requirement.
Penetration threshold.
Pattern high and low.
Nearby resistance reference.
Volume source and baseline, if used.
Indicator variables, if used.
Confirmation Rule 1–5.
Confirmation deadline.
Failure and invalidation rule.
Expiry horizon.
Favorable and adverse outcome thresholds.
Cost and fill assumptions.
Ambiguous-event rule.
Sample split and rule version.

Freeze the record before continuing the chart.

Evening Star Replay Worksheet

Chart identity

Instrument:
Venue/provider:
Price type:
Session:
Timezone:
Timeframe:
Adjustment method:
Bar state:

Prior advance

Advance rule version:
Lookback:
Start reference:
End reference:
Net change:
Pivot sequence:
Range break:
Qualified: yes / no / ambiguous

Candle 1

O1:
H1:
L1:
C1:
Range1:
Body1:
Body share1:
Bullish-body threshold met:

Candle 2

O2:
H2:
L2:
C2:
Range2:
Body2:
Body2 / Body1:
Body share2:
Bullish / bearish / doji:
Body gap:
Full-range gap:
Small-body threshold met:

Candle 3

O3:
H3:
L3:
C3:
Range3:
Body3:
Bearish requirement met:
First body midpoint:
Penetration:
Pattern completed:

Context

Pattern high:
Pattern low:
Resistance reference:
Distance to resistance:
Higher-timeframe state:
Volume source:
Volume ratios:
Optional indicators:

Confirmation and outcome

Confirmation rule:
Confirmation deadline:
Confirmed bar:
Confirmation price assumption:
Failure reference:
Expiry horizon:
Favorable threshold:
Adverse threshold:
First event:
MFE:
MAE:
Bars to event:
Costs:
Ambiguous event:
Final label:

Audit

Rule version:
Development / validation / out-of-sample:
Screenshot or record ID:
Rule changed after outcome: yes / no
Reason for exclusion:
Notes:

How to Practice With ChartMini

Use ChartMini to study the sequence without seeing the right side of the chart.

Open ChartMini replay.
Verify the chart identity and session assumptions.
Hide future candles.
Advance until a possible bullish first candle appears after a defined advance.
Record candles 1 and 2 without assuming an Evening Star will complete.
Reveal candle 3 and apply the frozen gap, body, and penetration rules.
Label the sequence completed, disqualified, or ambiguous.
Apply one selected confirmation rule.
Record confirmation, failure, expiry, and outcome without changing the original pattern definition.
Repeat across development, validation, and out-of-sample periods.

What ChartMini does not reproduce

ChartMini does not reproduce:

live bid/ask spread;
order-book depth;
queue position;
broker routing;
partial fills;
guaranteed stop or limit fills;
market impact;
every provider's session and adjustment method;
emotional responses to live profit and loss.

Replay is useful for classification discipline and decision records. It is not proof of live execution quality or profitability.

Common Evening Star Errors

Calling two candles an Evening Star

The sequence is not complete until the third candle closes and meets the chosen penetration rule.

Requiring a stock-style gap in every market

A strict daily equity gap and a no-gap continuous-market version create different samples. Label them separately.

Treating the third candle and next-candle confirmation as the same event

The third candle may complete the pattern. A later break or retest is additional confirmation.

Changing the midpoint rule after seeing the result

A midpoint, first-open, or pattern-low threshold must be selected before classification.

Treating a small middle body as known indecision

The candle records a small open-to-close distance. It does not reveal participant motives.

Assuming volume proves institutional selling

Volume shows recorded activity under a specific source. It does not identify the traders.

Drawing resistance after the reversal

The level must be frozen before later candles are revealed.

Using one fixed timeframe for every market

Timeframes compress different amounts of trading activity and produce different sample sizes, spreads, and noise.

Deleting patterns that do not confirm

Unconfirmed, expired, failed, and unresolved candidates are part of the evidence.

Treating replay fills as live fills

Historical OHLC data cannot guarantee executable prices or event order.

Practical Next Step

Create one Evening Star rule card containing:

chart identity;
prior-advance definition;
candle 1, 2, and 3 thresholds;
gap version;
penetration rule;
one confirmation rule;
failure and expiry rules;
outcome horizon;
cost assumptions.

Classify a fixed development sample without changing that card. Revise the rule only after the sample is complete, assign a new version number, and evaluate the new version on different data.

Frequently Asked Questions

What is an evening star candlestick pattern?

An evening star is a three-candle bearish reversal candidate considered after a defined price advance. A reference version uses a relatively large bullish first body, a smaller middle body, and a bearish third candle that closes into the first candle's body. The exact body, gap, penetration, trend, and confirmation rules should be written before examples are classified.

Does an evening star require a gap?

Traditional definitions often include a body gap between the first and second candles, but strict full-range gaps are uncommon in continuously traded markets. A test should state whether it requires a body gap, a full-range gap, a higher middle body without a gap, or no gap condition, and should not mix those versions after seeing the outcome.

How deep should the third evening star candle close?

A common reference rule requires the third candle to close below the midpoint of the first bullish body. Stricter versions may require a close below the first candle's open or below the three-candle pattern low. No threshold is universally best, so the selected penetration rule should remain fixed within a test.

How is an evening star different from a shooting star?

An evening star is a three-candle sequence whose third bearish candle completes or confirms the pattern. A shooting star is a single-candle shape with a small body near the lower part of its range and a long upper wick after an advance. They may appear in similar locations, but their geometry and confirmation records are different.

When does an evening star fail?

Failure depends on the version being tested. Examples include the third candle not meeting the required penetration, price closing above the pattern high after completion, no downside follow-through within a fixed number of bars, or a downside break that quickly closes back above the predefined reference. Failure and expiry should be defined before replay.

Can ChartMini prove that the evening star is profitable?

No. ChartMini can hide future candles and help you record pattern, confirmation, failure, and outcome rules bar by bar. It does not reproduce live spreads, slippage, liquidity, order priority, market impact, broker fills, or emotional decisions, so replay results do not prove live profitability.

Sources and Evidence Boundaries

StockCharts ChartSchool: Candlestick Bearish Reversal Patterns — common Evening Star sequence and bearish-confirmation description.
StockCharts ChartSchool: Candlesticks and Resistance — use of bearish reversal patterns as possible resistance context.
TradingView: Evening Star — Bearish — built-in three-candle definition, midpoint close, and configurable trend detection.
CME Group: Technical Analysis — OHLC construction and the distinction between objective chart data and subjective pattern interpretation.
CME Group: Chart Types — Candlestick, Line, Bar — candle body, wick, gaps, and timeframe construction.
Charles Schwab: How to Read Stock Charts and Trading Patterns — chart construction, pattern context, and past-performance limitations.
CFTC: Commodity Trading Systems Sold on the Internet — limitations of hypothetical results, fills, liquidity, spreads, commissions, and selected historical trades.

These sources describe educational conventions and risk boundaries. They do not establish a universal Evening Star success rate, optimal parameter set, or guaranteed trade outcome.

Evening Star Candlestick: 5 Confirmation Rules Beginners Can Test