Dark Cloud Cover: Exact Rules, Real Examples, and the Crypto Catch

Dark cloud cover is a rally's morning-after candle. A strong green session closes near its high, the next bar opens above that entire bar — beyond its upper wick — and then collapses, closing deep inside the green body it had gapped over. Japanese traders named it kabuse-sen (かぶせ線), the "covering line," and the image is exact: a cloud rolling over the top of an advance.

Most guides stop at the picture. This page gives you the precise rule Vike's engine runs, a measurement of how brutally rare that rule is on a market with no opening bell, five real detections from Binance 4-hour charts, and one that failed with numbers attached.

What dark cloud cover means

Read the two sessions as an auction. The first is owned by buyers, closing with strength. The second opens even higher — every holder is in profit at that print — and from that best possible start, sellers drive price down through more than half of the previous bar's body.

That gap is what separates this from an ordinary red bar. Whoever paid the opening price watched the market reject it immediately, and everyone who bought the prior session's close is underwater by the finish. That is a genuine transfer of control, compressed into one bar's round trip.

Depth matters too, and it is bounded on both sides. Too shallow a close changes nothing; too deep — past the green bar's open — and you are looking at a bearish engulfing instead. Dark cloud cover lives strictly between those fences.

How to identify dark cloud cover: the exact rules

Four conditions, in the order the engine tests them:

• The first bar is green — a real advance to cover.
• The second bar is red.
• That red bar opens strictly above the first bar's high — above the wick, not merely the body.
• It closes below the midpoint of the green body, yet above the green bar's open — deep, but not a full cover.

How Vike computes it. batch_dark_cloud_cover in vike-indicators reads three columns — open, high, close — and never touches a low. It stamps -100 when open > prior high, close < (prior open + prior close) / 2, and close > prior open all hold across a white-then-black pair. No body size, no shadow, no volume, no trend clause: the low of either bar could be anything.

The midpoint is the body's, not the range's.

The gap clause, measured

Here is what no textbook mentions: on a 24/7 venue, that third clause nearly abolishes the pattern.

A crypto bar opens wherever the last one closed — exactly, to the tick, 37% of the time in my scan of 480,125 daily bars across 376 Binance pairs, and within a hair otherwise. Opening above the previous high therefore requires a genuine micro-gap through the whole upper wick, which only happens when the book is thin and the tape jumps between bars.

Run the cascade on daily bars and the story is short. Of 122,485 green-then-red pairs, just 178 opened beyond the prior high, 105 of those closed past the midpoint, and 39 survived the final fence — one detection per 12,311 daily bars, none of them on a major. So the examples on this page come from the 4-hour scan, where 1,495,766 bars produced 596 detections with a median qualifying gap of 0.063%. The rule is identical; shorter bars just leave a jump less room to hide. The evening star page documents the same phenomenon from the other side — its gap clause is why both star pages run on 4-hour candles too.

Dark cloud cover vs piercing

Flip it vertically and you get the piercing pattern: a red bar, a gap below its low, and a green recovery through the body's midpoint that stops short of the open.

What matters is how the crate encodes that mirror. These are two separate functions, mirrored clause for clause — and the mirror swaps which price column each one loads. batch_dark_cloud_cover destructures open, high, close, leaving lows invisible to it; batch_piercing takes open, low, close and cannot see a high. Same construction as the morning and evening stars, and unlike the engulfing family, where one function simply flips its sign.

One consequence I keep leaning on: because the final clause is strict, no pair of bars can ever be both a dark cloud cover and a bearish engulfing. Close above the prior open and you are here; close at or below it and the engulfing detector claims the pair. The 105 midpoint-passing dailies in my scan split exactly that way — 39 stayed dark clouds, 66 covered fully and left the definition.

How to trade dark cloud cover

Treat the gap as information about eagerness, not a level. Sellers absorbing an above-the-high open is the signal; the open itself is usually just a spike print on crypto.

Wait for the follow-through bar. A close below the two-bar low is the confirmation that the reversal is real. The WLD detection above closed 98.4% of the way down the prior body — nearly an engulfing — and still needed the next bar to prove it before the 7.2% slide.

Set the invalidation above the red bar's open. That open is the highest price anyone paid in the pattern; if the market returns there and holds, the cloud thesis is dead. It is a tight stop by two-bar standards, which is this pattern's one structural kindness.

Respect the base rate. Of 528 graded 4-hour detections, 294 fell at least 2% within five bars and 234 rose instead — a 56% edge before any context filter. Real, but nothing you can trade on its own, so I skip every print that lands mid-range with no rally behind it.

Where dark cloud cover fails

Start with the failed chart above. PLUME printed a textbook dark cloud on 2026-06-17 — a 0.086% gap over the high, a close 62.5% into the body — and the market rose another 5.7% within five bars, straight through the pattern.

Failures here come in two flavours I have seen repeatedly. In a strong uptrend the gap-and-fade is often just profit-taking, and the trend resumes as if nothing happened — that is the PLUME case. And on the thin pairs where the qualifying gap is even possible, the same illiquidity that created the gap can produce a meaningless recovery bar right after it.

When I grade these setups now, the question is never whether the two bars are shaped correctly — the engine settles that — but whether the rally behind them was already stretched. A dark cloud after a vertical run into resistance is a trade; the same print three bars into a fresh breakout is usually noise.

Frequently asked questions

Is dark cloud cover bullish or bearish? Bearish, always — the engine emits -100 and nothing else. Its bullish mirror image is the piercing pattern, a separate function in the crate.

Does dark cloud cover need a gap? Under Vike's rule, yes: the red bar must open strictly above the prior bar's high. That is stricter than the "opens above the prior close" some books teach, and it is the clause that makes the pattern near-extinct on daily crypto — 39 detections in 480,125 bars.

Where must the second candle close? Past the midpoint of the first bar's body but above its open. Both fences are strict; exact ties fail.

What is the difference between dark cloud cover and bearish engulfing? The finish line. A bearish engulfing closes at or beyond the prior open; a dark cloud stops short of it. One boundary price, two mutually exclusive patterns — a pair can never be both.

This is educational material, not financial advice. Candlestick patterns describe probabilities, not certainties, and every real trade carries real risk — size positions so a wrong read costs little.

See the full candlestick pattern cheat sheet