Bearish Engulfing: The Number Points Wrong
A bearish engulfing pattern is a red candle whose body completely covers the previous green candle's body, read as sellers taking control. On this site's shared history it occurs 70 times and is followed by a higher close 58% of the time, which points the opposite way to what the pattern asserts.
How it works
Two candles. The first closes higher than it opened; the second closes lower than it opened, and its body covers the first one’s body entirely. Open above the previous close, close below the previous open.
It is the bullish engulfing pattern with the signs reversed, and the definitional questions are identical: bodies rather than full ranges, and a required context of a prior advance that screeners routinely ignore.
The story is that sellers took control in a single session and gave back everything the buyers had achieved the day before. Like every candlestick narrative, it is a plausible account constructed from four prices per bar and no information about who traded.
The measurement, and what to do with it
On this site’s shared 576-bar history, 70 bearish engulfing patterns appear — 12.2 per hundred bars, the most common pattern of any counted in this set.
Of the 69 that had a following bar, the next close was higher 58% of the time.
The base rate across all 575 transitions is 51%. So a pattern that claims sellers have taken control was followed by a higher close seven points more often than an average bar — the largest deviation of any pattern measured here, in the wrong direction.
This is not a discovery about markets, and it is important to say so directly. The series behind it is synthetic. It has no earnings, no participants and no memory of its own past. A pattern cannot have predictive power on it, because there is nothing to predict from.
Which is exactly why the number is worth publishing. A seven-point deviation on 69 events emerged from a process that guarantees no effect exists. That is what seven points on seventy observations is worth: nothing. And it is the same order of magnitude as the deviations reported, with confidence, in a great deal of published candlestick research.
There is also a mechanical explanation available. A bearish engulfing bar is by definition a large down bar. Large down bars are followed by bounces at a slightly elevated rate in most price series, including random ones, simply because a big move in either direction is often partially retraced. The pattern is selecting for bar size, and the follow-up statistic is picking up that selection rather than anything about sellers.
In practice: what the pattern actually selects for
Volume is the only input that could turn the shape into evidence. A large down bar on the heaviest participation in months is a genuinely different event from the same bar on nothing, and the candle shape treats them identically.
A gap up that closes red satisfies the test automatically. Opening above yesterday’s close is half the definition, and the closed market supplied it. Nobody was overwhelmed.
Aggregate two daily bars into one weekly bar and the pattern ceases to exist. It is a relationship between adjacent bars, and bar boundaries are a setting rather than a property of the market.
Seventy signals at 2% of a typical bar’s range each is the largest cost of any pattern in this set, in round-trip costs on this history — and if the trade is a short, borrow costs sit on top of that.
And the order book has no record of yesterday’s range. The engulfing is a relationship between two drawings on your screen. There is no mechanism by which one bar covering another changes what the next participant does.
What a bearish engulfing is not
It is not a top. It is a common two-bar shape, occurring 12 times per hundred bars here, and there were nowhere near that many tops.
It is not a stronger signal than its bullish mirror. They are the same construction. If one works, the argument has to explain why the other does not, and no such argument exists.
It is not an outside bar. Outside bars compare full ranges; engulfing conventionally compares bodies. The two overlap and are not the same test.
And it is not confirmation of a downtrend. By definition the preceding bar was green, so the pattern requires an up bar to exist at all. That makes it a bar-to-bar reversal, not a statement about trend.
When it fails
In a range they print at every high. Price reaches the ceiling, turns, and the down bar covers the up bar before it. That is the definition being satisfied by ordinary oscillation, and it accounts for a large share of the 70 occurrences here.
The second failure is the entry. Selling the close of the engulfing bar means selling the low of a large bar, with the logical stop above its high. The pattern qualifies by being big, and being big is what makes the entry poor.
A third is the survivorship in the teaching material. Every article on this pattern shows examples that preceded declines. The ones that preceded nothing are not memorable and never get published, so the visible record is filtered before you see it.
A fourth is stacking it with an oscillator. The relative strength index, RSI, or the moving average convergence divergence indicator, MACD, alongside an engulfing candle looks like independent agreement and is two readings of the same recent closes, which the confluence page treats in detail.
And a fifth is trusting a percentage without an occurrence count. This page’s own figure is the demonstration: 58% on 69 events, from a series where no effect can exist.
The original data
70 bearish engulfing patterns in 576 bars — 12.2 per hundred, the most frequent pattern counted here —
followed by a higher close 58% of the time on the 69 with a following bar, against a 51% base rate across
all 575 transitions. Everything is in research/series-measurements.json, produced by
site/measure_series.py.
Set the two engulfing patterns side by side and the lesson finishes itself. The bullish version landed exactly on the base rate across 63 events; the bearish version landed seven points on the wrong side across 69. Same construction, same data, opposite results — from a series that cannot contain a signal. A method that produces contradictory findings on data known to have no effect is a method that will produce confident findings on data that has none either. That is the case for demanding sample sizes and base rates before believing anything about candlestick patterns, and it is why both figures are published here rather than only the flattering one.
Related
Bullish engulfing is the mirror pattern with the null result on the same data. Candlestick patterns is the parent page for the naming system. And reversals treats the event this pattern claims to mark.
Finding a number that pointed the opposite way to the textbook was the moment this whole exercise became worth doing. It is not that bearish engulfing candles are secretly bullish - it is that a synthetic series with no participants in it produced a seven-point deviation, which tells you what seven points is worth.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.