Candlestick charting was developed for Japanese rice markets and refined for centuries on instruments that close every day. Crypto does not close. That difference is not cosmetic, and it quietly breaks a specific subset of the classic patterns while leaving most of them intact.
Here is which is which, and why.
What a session boundary actually provides
On a stock, a trading day is a container. It has an open that follows hours of accumulated news, a close that people position for, and a gap between the two that is a real event with real information in it. Several candlestick patterns are built on that structure rather than on price alone.
Crypto has none of it. A “daily” candle on BTC/USD is a 24-hour slice whose boundary is an arbitrary line drawn at midnight UTC, or whenever your exchange decides. Nothing happens at that line. Nobody positions for it. Price at 23:59 becomes price at 00:00 with no discontinuity.
That single fact determines everything below.
The patterns that break
Anything that requires a gap. Dark cloud cover is the clearest case: its definition requires the second candle to open above the previous candle’s high, then close below the midpoint of its body. On a continuously traded instrument, one candle’s open is essentially the previous candle’s close, so that opening gap almost never occurs. What crypto traders usually label dark cloud cover is a strong down candle following a strong up candle, which is a different and much weaker observation.
The same caution applies to the gapping variants of the star patterns. The classical morning star and evening star descriptions have the middle candle gapping away from the first. On crypto it rarely does. The patterns still form in their non-gapping version, and they still describe a genuine stall, but you are looking at a weaker instance than the literature is describing, and you should weigh it as one.
Anything that leans on an exact price match. The tweezer patterns rely on two candles sharing nearly the same high or low. In a market with no session boundaries and often thinner order books, price prints at far more distinct levels, and two “matching” extremes within a few ticks are more likely to be coincidence than a defended level. The pattern is not wrong on crypto; it is just much easier to see where none exists.
Exact dojis, on thin pairs. A doji requires the open and close to be effectively equal. On a liquid pair like BTC/USD that is a real standoff. On a low-volume altcoin on a 15-minute chart, it frequently means almost nothing traded, which is a liquidity artefact wearing a pattern’s clothes. This is not unique to crypto, but crypto has far more thin instruments than equity markets do.
The patterns that survive intact
Everything whose meaning comes from within-candle behaviour rather than between-candle structure works exactly as it always did, because it only ever needed price to move.
- The hammer and shooting star: a long wick is a rejection of a price, and rejection does not care whether a bell rang.
- The engulfing patterns: one body covering another is a statement about range, not about opens.
- Three white soldiers and three black crows: each candle opening inside the last one’s body and closing near its own extreme is, if anything, more meaningful on a continuous market, because the “opens inside the previous body” condition is automatic there and the information is entirely in the closes.
- Spinning tops and indecision generally: contested sessions are contested sessions.
That is the large majority of what is worth knowing. The honest summary is that crypto breaks the gap-dependent minority and leaves the price-action majority alone.
The thing that actually changes more than the patterns
Not the shapes. The timeframe you should be reading them on.
A daily candle on a stock represents one full cycle of human attention: a session that opened, ran, and closed, with participants who mostly went home. A daily candle on crypto represents 24 hours spanning three continents’ waking hours, with no natural rhythm inside it. The two are not comparable units even though the chart draws them the same way.
In practice this means crypto’s 4-hour chart is closer in character to a stock’s daily than crypto’s daily is, because it captures something nearer to one region’s active period. If you learned patterns on equities and they feel noisy on crypto, that mismatch is usually the reason rather than the patterns failing.
The general rule still holds either way: the longer the timeframe, the more weight a pattern carries, because more accumulated opinion went into forming it. The far end of that scale has its own problems, covered in candlestick patterns for intraday trading.
What to actually do with this
Three adjustments, and they are small:
- Discount the gap-dependent patterns rather than forcing them. If a pattern’s definition includes an opening gap and there is no gap, it is not that pattern.
- Tighten what counts as a match on tweezers and dojis, or stop treating them as standalone evidence and use them as location markers.
- Shift your reading timeframe up relative to where you would read equities, and stop expecting the daily to feel like a daily. The forex version of the same question is covered in forex candlestick patterns.
None of this changes the discipline that matters most, which is the same in every market. Name what the pattern requires before it means anything, and name the price that would prove the read wrong. The market being open at 3am does not make that part optional.
For the full set of patterns with their context requirements and invalidation levels drawn out, see the candlestick pattern library. If you want the underlying mechanics first, how to read a stock chart covers what the four prices in a candle record and why timeframe changes the picture.
Sources
Every figure above was checked against these before publishing.
- Bulkowski, Dark Cloud Cover
- Bulkowski, Morning Star
- Bulkowski, Tweezers Top
- Bulkowski, Three Black Crows
- Investopedia, Dark Cloud Cover
- CME Group, Trading Hours