Bullish Engulfing: Exactly the Base Rate
A bullish engulfing pattern is a green candle whose body completely covers the previous red candle's body, read as buyers overwhelming sellers. On this site's shared history it occurs 63 times and is followed by a higher close in 51% of cases, which is identical to the rate for any bar.
How it works
Two candles. The first closes lower than it opened; the second closes higher than it opened, and its body covers the first one’s body entirely. Open below the previous close, close above the previous open.
Bodies, not wicks — that is the standard test and it is worth stating because a good deal of confusion comes from sources that include the wicks and therefore count far fewer occurrences.
The story is a reversal of control. Yesterday sellers finished ahead; today buyers not only won but covered the whole of yesterday’s ground. It is a clean narrative and it maps onto the picture exactly.
The count, against the only comparison that matters
On this site’s shared 576-bar history, 63 bullish engulfing patterns appear — 10.9 per hundred bars. That makes it one of the most common named patterns there is, which is the first thing worth knowing: it is not a rare event that marks turning points, it is something that happens roughly once a fortnight on a daily chart.
Of those 63, the next bar closed higher 51% of the time.
Across all 575 transitions in the same data, the next close was higher 51% of the time. The pattern and the base rate are identical to the nearest percentage point.
On a sample of 63 — the largest of any pattern counted here — the result is exactly nothing. That is a more informative outcome than a small edge would have been, because 63 events is enough to notice a large effect if one existed, and none did.
Two caveats, stated so the figure is not over-read. This history is synthetic and has no participants, so it cannot show a real behavioural effect if one exists. And 63 is still a modest sample. What the number does establish is that on a series with no mechanism, this pattern produces no signal — which is the correct null result, and the thing to compare a real measurement against.
In practice: the structural problem with the entry
To engulf the previous body, the second candle has to be large. That is not incidental — it is the qualifying condition. So the bar that confirms the pattern is, by construction, one of the bigger bars on the chart.
Which means the entry is at the top of a big bar and the logical stop is at the bottom of it. The distance between them is wide precisely because the pattern qualified, so every valid bullish engulfing offers a worse risk-per-unit than an ordinary entry would.
That is a real, structural cost and it is almost never mentioned. It applies whether or not the pattern has any predictive value: the better the pattern looks, the larger the second bar, and the worse the entry.
Volume is what separates a genuine change of control from a large bar on nothing. An engulfing candle on the heaviest participation of the month is a different event from one on a quiet holiday session, and the shape does not distinguish them.
A gap down that closes green satisfies the definition mechanically. Opening below yesterday’s close is half the test, and the market being shut supplied it. No sellers were overwhelmed; the session simply started lower.
Change the aggregation and the pattern evaporates. Two daily candles become part of one weekly candle with an ordinary body. The pattern exists at the bar length you happen to be looking at, which is a choice you made.
63 signals at 2% of a typical bar’s range each is a substantial bill in round-trip costs on this history — and against a measured edge of zero, it is the whole result.
And nothing engulfed anything. Two bars have a geometric relationship on a chart. The order book contains no memory of yesterday’s range and no mechanism by which today’s bar covering it changes what anybody does next.
What a bullish engulfing is not
It is not a reversal. It is a two-bar shape that appears at some reversals and at a great many continuations, retracements and ordinary oscillations.
It is not the same as an outside bar. An outside bar requires the whole range — high and low — to exceed the previous bar’s. Engulfing conventionally tests bodies only, so an engulfing candle need not be an outside bar and vice versa.
It is not stronger for being bigger. A larger second bar makes the pattern more visually convincing and the entry worse, which is close to the opposite relationship from the one people assume.
And it is not two pieces of evidence. It is one shape defined across two bars, which is a single observation about a short stretch of price.
When it fails
In a range they occur at every turn. Price falls to the floor, turns, prints a large green bar that covers the previous red one — that is the definition being met by ordinary oscillation. At 10.9 occurrences per hundred bars, most of them are this.
The second failure is the entry price. Buying the close of the engulfing bar means buying the high of a large bar. Waiting for a pullback improves the entry and frequently means the trade never happens, which is the trade-off nobody resolves.
A third is the definition ambiguity. Bodies or full ranges? Must the second body exceed the first at both ends, or only at one? Different sources answer differently, so occurrence counts and published statistics are not comparable across articles.
A fourth is ignoring what preceded it. The pattern is defined as appearing after a decline. Halfway up an advance, the same shape is a continuation bar wearing a reversal name.
And a fifth is treating a common event as a rare one. A signal that fires 63 times in 576 bars cannot be marking major turning points, because there were not 63 major turning points. The frequency alone tells you what kind of thing it is.
The original data
63 bullish engulfing patterns in 576 bars, followed by a higher close 51% of the time, against a 51%
base rate across all 575 transitions. The counts and the exact test used are in
research/series-measurements.json, produced by site/measure_series.py.
A null result on the largest sample in the set is the most useful number on this page. Every other pattern measured here landed a few points from baseline on 25 to 31 events, where noise dominates. This one had 63 events and landed exactly on it. When you next read a candlestick statistic, look for the occurrence count: the patterns common enough to measure tend to show nothing, and the ones showing something tend to be too rare to measure. That trade-off is structural, and it is why this corner of technical analysis has so much written about it and so little settled.
Related
Bearish engulfing is the mirror pattern, and its number on this data points the wrong way. Candlestick patterns is the parent page for the vocabulary. And reversals covers the event this shape claims to identify.
Engulfing candles were the pattern I found easiest to spot and hardest to trade, and the reason turned out to be structural rather than psychological. The bar that qualifies the pattern is by definition a big one, so by the time it closes you are entering at the top of it with your stop at the bottom - which is a worse trade than the picture suggests.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.