How to Read Candlestick Charts — Properly, Including the Limitations
A candle encodes four numbers. Most of what is taught about candlestick patterns is elaboration on top of those four numbers, with the context that gives them meaning left out. Here is the version with the context included.
What a candle actually encodes
Each candle records four numbers for its period: open, high, low and close. The body spans open to close; the wicks extend to the extremes.
That is the entire information content. Everything taught about candlesticks is interpretation layered on those four numbers — which does not make it worthless, but does mean that any claim about what a candle "tells" you is an inference, not a reading.
What the arrangement genuinely shows is the relationship between travel and settlement: how far price went during the period, and where it ended relative to that. A period that ranged widely and closed near its open describes a different balance of activity from one that moved directly and closed at its extreme.
Reading a single candle
| Shape | What it describes | What it does not tell you |
|---|---|---|
| Long body, short wicks | Directional period — price moved and held | Whether it continues |
| Small body, long wicks both sides | Indecision — both directions attempted and rejected | Which side eventually wins |
| Long lower wick, close near high | Lower prices were attempted and rejected within the period | Whether the rejection holds |
| Long upper wick, close near low | Higher prices were attempted and rejected | Whether sellers persist |
| No body — open equals close | Perfect balance over the period | Almost nothing on its own |
Note the third column. Each shape is a description of what happened, not a forecast. Treating a long lower wick as a buy signal skips the step where you establish whether that rejection occurred anywhere meaningful.
The patterns that carry information
Of the dozens catalogued, a small number describe genuinely informative structures — and each is informative because of what it says about a failed attempt, not because of its shape:
- Engulfing. A candle whose range fully covers the previous one, closing in the opposite direction. Describes a decisive reversal of the prior period's balance. Informative at a defined level; noise mid-range.
- Pin bar / long-wick rejection. A sharp excursion beyond a level, fully rejected by the close. Describes an attempt that failed. Most useful at a prior-day extreme, VWAP, or a well-defined boundary.
- Inside bar. A period contained entirely within the previous one — a compression, describing reduced participation. Frequently precedes an expansion, without indicating direction.
- Doji at an extreme. Balance appearing after a directional run. Describes the run losing its one-sidedness. Says nothing about what replaces it.
- Marubozu / wide-range close at extreme. A period of sustained one-directional pressure, most informative when accompanied by expanding volume.
Everything else in the standard catalogue is largely a variation on these, distinguished by naming rather than by information content.
Why context beats the pattern
The same candle shape means different things in different places. Three context variables do most of the work:
- 01LocationAt a prior-day high, a multi-day range boundary, or VWAP — the pattern describes a specific failure. In the middle of a range it describes nothing.
- 02VolumeA rejection on heavy participation is a different event from the same shape on thin volume. This single filter removes a large share of false signals.
- 03Prevailing trendA reversal shape against a strong higher-timeframe trend is a much weaker proposition than the same shape in a range or at a trend's exhaustion.
You are not looking for patterns. You are looking for evidence that an attempt at a meaningful level failed. Candles are a convenient way to see that evidence — the level and the participation are what make it evidence.
Timeframe changes everything
A one-minute candle records a few moments of order flow. A daily candle summarises an entire session of participation across all market participants. They are not the same object, and the same shape does not carry the same weight.
- Lower timeframes contain proportionally more noise. Pattern signals on very short intervals fire constantly and mean correspondingly less.
- Higher-timeframe context should filter lower-timeframe entries. A long setup on a 5-minute chart inside a clear daily downtrend is fighting the current.
- Costs are a bigger share of a smaller move. Trading small patterns on short timeframes means every trade's edge is thinner relative to transaction costs.
The evidence problem
Candlestick literature is full of confident win-rate claims that do not survive examination. Three recurring issues, worth knowing before relying on any published figure:
- Definitional looseness. How long must a wick be to count as a rejection? How close to the extreme must the close be? Different definitions produce very different results from the same data, and most sources do not state theirs precisely enough to reproduce.
- Illustration bias. Teaching material shows examples where the pattern worked. That is pedagogically natural and evidentially worthless.
- No cost or exit specification. A pattern's "success rate" depends entirely on what counts as success — how far, over how long, net of what costs. Without those, the number is not measuring anything specific.
The practical conclusion is not that price action is useless. It is that any pattern you intend to trade should be defined precisely enough for you to test yourself, with your own exit rules and your own costs.
Using candles in a rules framework
- Define the pattern numerically. Wick-to-body ratio, close position within the range, minimum candle range relative to ATR. If it cannot be specified, it cannot be tested.
- Require a location condition. The pattern only counts within a defined distance of a reference level.
- Require participation confirmation. Volume above a defined threshold relative to its recent average.
- Take invalidation from the candle itself. This is candlesticks' most practical contribution — the wick's extreme is a natural, unambiguous invalidation level.
- Size from that distance. A long wick means a wider stop, which means a smaller position — not the same position with more risk.
- Test on your own data with your own costs before trusting any published statistic.
EqtPulse's research is rules-based, and where price structure contributes to a signal it is specified numerically rather than visually — because a pattern that cannot be defined cannot be tested, and a pattern that has not been tested is an opinion with a chart attached.
Frequently asked questions
What do candlestick charts tell you?
Four prices per period — open, high, low and close — displayed so the relationship between them is visible at a glance. That relationship indicates where price travelled during the period and where it settled relative to that travel.
That is genuinely useful information. What candles do not encode is why the movement happened or what happens next.
Are candlestick patterns reliable?
Individually, far less than commonly taught. Most published pattern win rates come from studies with loose definitions, no cost assumptions, and substantial selection bias in which examples get shown.
Patterns carry more information when treated as one input alongside location, volume and trend context than when treated as standalone signals.
Which candlestick pattern is most reliable?
Rather than a single pattern, the most consistently informative structures are those describing failed attempts at a level — a long wick rejecting a level on high volume, or an engulfing candle at a well-defined boundary. What makes them informative is the location, not the shape.
The same shape in the middle of a range means very little.
Do candlestick patterns work on all timeframes?
The arithmetic is identical, but the significance is not. A one-minute candle records a few moments of order flow and is dominated by noise; a daily candle summarises a full session of participation.
The same pattern shape carries far more information on a higher timeframe, simply because more participation went into forming it.
Can I trade using only candlestick patterns?
You can identify setups from price action alone, and many systematic frameworks do. What you cannot do without the rest of a framework is manage the outcome — position sizing, invalidation, cost accounting and regime filtering decide profitability regardless of how the setup was found.