Key takeaways
- A candle summarises open, high, low and close; its shape tells a short story about one period.
- Meaning depends on location, trend and volatility, not on the name alone.
- Reliability is modest; small samples and hindsight exaggerate it.
- The strongest use is defining a clear invalidation point and risk distance.
Candlestick charts pack four prices into one shape, and a library of named patterns has grown around them: hammers, dojis, engulfing bars, stars. The names are memorable, which is part of the problem. A memorable name can feel like a prediction. This guide explains what each shape records and how much weight a pattern can honestly carry.
What a candle records
Each candle summarises one period, such as an hour or a day, using four prices: open, high, low and close. The thick part is the body, running from open to close. The thin lines above and below are the wicks (also called shadows), reaching to the high and the low. A body that closes above its open is usually drawn in one colour and a lower close in another.
Read this way, a candle is a short story. A long body with tiny wicks says price moved one way and stayed there. A small body with long wicks says price travelled and was pushed back. The pattern names are shorthand for a few recurring stories.
Common patterns and what they describe
| Pattern | Shape | Story it tells | Context it needs |
|---|---|---|---|
| Doji | Open and close almost equal, wicks either side | Buyers and sellers balanced for the period | Meaning depends on what came before; in a quiet range it is routine |
| Hammer | Small body near the top, lower wick about twice the body or more | Sellers pushed price down, buyers took it back | After a decline, near a prior low or a support zone |
| Shooting star | Small body near the bottom, long upper wick | Buyers pushed up and were rejected | After an advance, near a prior high or resistance zone |
| Engulfing | Second body fully covers the previous body | One side overpowered the other for the period | After a move, ideally at a notable level |
| Inside bar | Whole range sits inside the previous bar | Pause and compression | Used as a setup for a break in either direction |
| Morning or evening star | Three candles: strong move, small candle, strong reversal | Momentum fading and turning | Needs a clear prior trend to be meaningful |
Shapes such as the hanging man look identical to a hammer and are distinguished only by where they appear, after an advance rather than after a decline. That tells you something important: the candle alone does not carry the meaning. Its location does.
How reliable are they?
Honest answer: less than their reputation suggests. Research on the profitability of candlestick patterns has produced mixed results, and many studies find the apparent edge to be small or to vanish once costs are included. This does not mean patterns are useless. It means that treating a name as a forecast is a mistake.
There are two practical reasons why a pattern is weaker than it looks. First, hindsight: on a finished chart, the hammers that preceded rebounds are easy to spot, while those that were followed by further declines are quietly forgotten. Second, small samples. The following illustration shows how little a modest hit rate proves.
Illustrative figures only: 52 of 100 occurrences were followed by a move in the expected direction.
A 52% result sounds like an edge. But the statistical noise on 100 coin-flip-style outcomes is about 5 percentage points (the square root of 0.25 ÷ 100 is 0.05). A reading of 52% sits only 0.4 of a standard error above 50%, which is entirely consistent with chance. Add the fact that you probably tested several patterns and keep only the best, and the evidence gets weaker still.
Using candles with context
The most defensible use of candlesticks is descriptive. They help you read what happened at a level you identified for other reasons, and they give you a precise point to define risk. Context filters that traders commonly add include the following.
- Location: is the candle at a meaningful support or resistance zone, or in the middle of nowhere?
- Trend: is the pattern with or against the higher-timeframe direction?
- Size: is the range large or small relative to recent candles, for example against the ATR?
- Confirmation: did the next candle follow through, or did the pattern fail immediately?
- Timeframe: patterns on higher timeframes involve more activity and are generally less random than those on one-minute charts.
Pattern geometry as a risk tool
A pattern's real practical value is that it hands you a natural invalidation point. Take a hypothetical hammer on a currency chart: high 1.2540, low 1.2500, close 1.2535. A trader who plans to stop out beyond the low, with a 5-pip buffer at 1.2495, and who enters near the close, risks about 40 pips. If the account risk per trade is 100 and one pip is worth 10 per standard lot, the position size is 100 ÷ (40 × 10) = 0.25 lots. The numbers are illustrative and are not a trade suggestion; they show how the candle's shape converts into a size.
Notice what the pattern did and did not do. It supplied a stop level and a risk distance. It did not supply a reason to expect a particular outcome.
Judging a pattern before relying on it
- Define it precisely. Write exact rules, such as the wick-to-body ratio and what counts as a prior decline. Vague definitions can be bent to fit.
- Collect all occurrences. Gather every instance in your data, not just the memorable ones, and record the result in R.Tip: Daily candle timing differs between platforms because brokers use different server times for the daily close, so the same pattern may not appear on two feeds.
- Add costs. Subtract spread and commission, and count slippage on stops.
- Check sample size. Be cautious with fewer than around a hundred cases, and check different periods and markets.
- Compare with a plain baseline. Ask whether the pattern beats simply entering at random times with the same stop and target.
Frequently asked questions
Which candlestick pattern is the most reliable?
No single pattern stands out consistently across markets and periods. Reliability varies with context, timeframe and costs, so test any pattern you plan to use on your own instruments.
Do I need to memorise dozens of patterns?
No. Understanding what long bodies, long wicks and small bodies mean is more useful than memorising names. A few shapes read well in context are enough.
Are candlesticks better than bar charts?
They contain the same four prices. Candles make the relationship between open and close easier to see, which is a matter of preference and not of information.
Should I wait for the candle to close before acting?
A candle is not final until its period ends, and a hammer can turn into something else by the close. Waiting avoids acting on a shape that has not formed yet, at the cost of a worse entry price.