Bearish Harami — The \"Top Signal\" That Usually Isn't

There's a pattern in the textbooks that "signals a top": after a large white candle of gains, a small black one appears, tucked inside its body. Bearish harami. The problem is that when its behavior was measured on 4.7 million candles, something inconvenient came out: more often than a top, the pattern preceded... further gains — in 53% of cases the market simply continued the trend. It's one of the cleanest examples of a gap between candlestick theory and measurement. And one of the best lessons in why you need to check the numbers before you put money on them.

What the Pattern Looks Like

Bearish Harami is a two-candle setup at the top of an upward move — the mirror image of bullish harami:

• Trend before the pattern: up. The pattern only makes sense if there's something to reverse.
• Candle 1: a tall white (bullish) candle. Buyers at full strength, a clear body.
• Candle 2: a small black (bearish) candle whose open and close fit inside candle 1's body. Wicks are ignored — only bodies count.
• The bodies can't fully overlap. The tops or bottoms of the bodies may sit at the same level, but not both at once.

The textbook narrative: after a strong rally, buyers suddenly can't push price higher, and the small black candle is the "first crack" in the trend. The problem with this narrative is the same as with any harami — a small inside candle doesn't show the other side taking control. It shows a pause. And a pause in an uptrend is statistically more often a rest before the next move up than the start of a bear market. That's exactly what the measurement found.

On crypto the pattern occurs naturally and often — it needs no gaps, so a 24/7 market doesn't distort it in any way. On BTC/ETH, after a strong green D1 candle, a small inside red candle is almost routine — which by itself should be a warning: a signal you see every week is rarely an edge.

What the Numbers Say

Results from Bulkowski's tests (~4.7 million daily candles, US stocks) — and this is the section worth opening the article for:

• Behavior measured: gains continued in 53% of cases. The theory says "bearish reversal." The measurement says: more often, price closes above the pattern's high and keeps climbing. The uptrend leading into the pattern helps it... continue that trend. Bulkowski calls this "nearly random" — but random with a slight lean against the textbook.
• Overall performance rank: 72/103. Weak. Even when a breakout does happen, the move that follows is among the poorest — across all combinations of bull/bear market and breakout direction, the 10-day price change doesn't even reach half of the 6% threshold Bulkowski considers a good result.
• Frequency: 26/103. A common pattern — you'll find plenty of it in any price history. The author of the measurements sums it up dryly: whether it's worth hunting for, he leaves up to you.
• Best average 10-day move: -4.01% (bear market, downside breakout) — 50th place out of 103. Best price-target hit rate: 64% (bull market, downside breakout).

Honest synthesis: bearish harami as a standalone top signal is exactly what a coin flip with a slightly loaded coin looks like — loaded against you. The pattern's informational value is limited to the sentence "upward momentum has paused." Standard disclaimer: the measurement covers US stocks on the daily timeframe; on crypto (24/7, different volatility) treat the percentages as indicative — though the fact that a small candle after a large green one is a daily occurrence on BTC arguably reinforces the conclusions rather than weakening them.

How to Trade It (and How Not To)

How NOT to trade it: shorting "because there's a harami at the top." This trade has negative expected value already at the level of the raw direction statistic: in 53% of cases the market keeps climbing, taking out your stop. On top of that, rank 72 means even the winning shorts delivered a small move on average. In short: trading this pattern by the book is a systematic way to give away money.

Scenario 1 — harami as a warning signal for a long. The only use that doesn't require predictive power the pattern doesn't have. You're holding a long after a strong rally, the market prints a harami at resistance — that's not a reason to close the position, but a reasonable moment to raise the stop or take partial profit. If the pattern turns out to be a pause (statistically more likely), the position keeps running. If it turns out to be a top, your gains are protected.

Scenario 2 — wait for three inside down. Bearish harami confirmed by a third candle closing below the pattern's low is a separately catalogued pattern, three inside down: a reversal in 60% of cases, rank 56/103. Still nothing spectacular, but the gap between 47% and 60% success is the difference between trading against the statistic and trading with it. Practical takeaway: you never trade the harami alone — you trade (at most) its confirmation.

Scenario 3 — the best setup per the measurement: a return to a downtrend. Bulkowski points out directly: bearish harami works best not at the top of a bull run, but at the top of an upward correction within an ongoing downtrend. A downside breakout then doesn't reverse anything — it rejoins an existing trend. On crypto: a bear market on the D1, a bounce up into a resistance zone, a harami there, then a close below the pattern's low — only that full sequence makes sense as a short. Extra filter from the data: a pattern at the upper edge of a trend channel more often breaks out downward.

Stop loss and target. For a short after confirmation: stop above the pattern's high (including wicks), target at the nearest support, partial realization. The pattern is small, so the stop ends up close — don't turn that into an oversized position, because losing streaks are baked into this statistic.

Myth vs Measurement

Myth: "Bearish harami is a reversal signal at the top." Measurement: gains continue in 53% of cases. The pattern fails more often than it works — this is literally a pattern with a negative record against its own theory.

Myth: "A small black candle after a big white one shows sellers entering the market." Measurement and logic: a small inside candle only shows the absence of continuation in a single session. Sellers who actually enter the market look like bearish engulfing (79% reversal) or dark cloud cover (60%) — large black bodies, not small ones. The size of the attacking candle matters, and the measurement consistently shows it.

Myth: "If a pattern is in every textbook, it must work." Measurement: rank 72/103 and a success rate below the theoretical baseline. Candlestick manuals were built on rice markets and Japanese stocks decades ago — nobody counted their authors' accuracy on millions of candles. Bulkowski did. The result is above.

A Sample Scenario on BTC

BTC in a daily downtrend bounces up for a week and reaches a resistance zone — a previous support level that flipped polarity after breaking. There it prints a large green candle, followed by a small red one inside its body. This is exactly the preferred setup from the measurement: bearish harami at the top of a correction in a bear market. You still don't enter — you set an alert at the pattern's low. A D1 close below it gives you a three inside down aligned with the dominant trend: a short with a stop above the pattern's high and a target at the last low. If instead price closes above the harami's high — the statistic just played out as modeled, and you're glad you didn't enter earlier.

Quick checklist:

• Does the second candle's body sit entirely inside the first candle's body (wicks ignored)?
• Is the pattern at real resistance, ideally as the top of a correction within a downtrend?
• Are you waiting for a close below the pattern's low (three inside down) instead of shorting immediately?
• If you're holding a long, did you treat the pattern as a position-management signal, not a panic exit?
• Is the stop above the pattern's high, with position size calculated from the stop?
• Do you remember that, on base rates, the pattern continues gains more often than it reverses them?

Bearish harami is a memento pattern: between "that's what they teach" and "that's how it is" there can be a chasm exactly where you place your stop. Whoever trades the textbook picture is trading against measured reality. Whoever knows the measurement uses this candle for what it's actually good for — a reminder to manage the position, and a component of a setup, never a setup on its own.

FAQ

Rafał — Krypto Bez Ściemy

Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.

kryptobezsciemy.plInstagramYouTubeKlub Zero