Abandoned Baby — The Three-Candle Island Reversal

In the performance ranking of 103 candlestick patterns, the bullish Abandoned Baby sits at 9th place, and its best measured variant held 1st — literally first — position in the 10-day move ranking. Sounds like the Holy Grail? Here's the other side of the coin: out of 4.7 million candles tested, Bulkowski found 293 instances of this pattern. And if you trade crypto on spot, the number for you is a flat zero — because this pattern requires price gaps, which a 24/7 market simply doesn't produce. So this article is half about candles, and half about why you need to read a study's fine print.

What the Pattern Looks Like

Abandoned Baby is a three-candle reversal pattern built around an island doji — a candle completely separated from its neighbors by gaps. The bullish variant:

• Trend before the pattern: downtrend.
• Candle 1: bearish (black) — shape and size don't matter.
• Candle 2: a doji that gaps DOWN on the open, so its upper wick stays below the lower wicks of both neighboring candles. This is the defining condition: not a gap between bodies, but full separation, wicks included. The doji hangs alone below the market — hence the name.
• Candle 3: bullish (white), gaps UP on the open — its lower wick stays above the doji's high.

The bearish variant is the mirror image: an uptrend, a white candle, a doji gapping up (its lower wick above the white candle's high), then a black candle gapping down (its upper wick below the doji's low).

The story is dramatic: the market opens with a gap in panic (or euphoria), can't pick a direction all session (the doji), and the next day opens with a gap in the opposite direction and takes off. The last buyers (sellers) who entered on the doji were "abandoned" on an island — anyone who entered there is instantly underwater.

And here's the no-hype disclosure promised in the intro: on crypto spot, this pattern practically never occurs. A 24/7 market doesn't know gaps — every candle opens exactly where the previous one closed, so the full-separation condition for the doji can't be met (barring extreme liquidity failures on thin altcoins, which you don't want to trade anyway). If a pattern scanner shows you an "abandoned baby" on BTC/USDT, it's almost certainly a Morning Doji Star or Evening Doji Star without the gap — a related pattern with its own, weaker statistics. Look for the real Abandoned Baby where sessions and gaps exist: stocks, indices, CME Bitcoin futures with their weekend gaps.

What the Numbers Show

Results from Bulkowski's tests (~4.7 million daily candles, US stocks):

• Bullish variant: reversal in 70% of cases, overall performance rank 9/103. On paper — top of the catalog.
• Bearish variant: reversal in 69% of cases, but rank only 64/103. Reverses often, but the move after the reversal is weak — the study's own author calls this performance poor and warns not to expect a long trend.
• Frequency: 92/103 (bullish) and 96/103 (bearish). Extreme rarity: 293 bullish cases in 4.7 million candles. Bulkowski jokes he checked "every adoption agency in the state."
• Best average 10-day move (bullish): -10.31% after a downside breakout in a bear market — ranked 1st out of 103 patterns. Except that number rests on... 14 cases. The author himself writes plainly: the result is unreliable and will almost certainly change.

This is the real lesson of Abandoned Baby. The pattern has stellar headline statistics and, at the same time, a sample size so small those statistics barely mean anything. 293 cases out of 4.7 million candles is 0.006% — and the flashiest numbers (that rank 1!) come from subsets with a dozen or so observations. In bear markets, all results for both variants rest on fewer than 20 samples. Anyone selling you "the #1 ranked pattern" is quoting noise.

The market caveat this time is double: first, the standard one — US stocks, daily timeframe, different percentages than crypto. Second, the fundamental one — on crypto spot, the pattern doesn't occur at all, so there's nothing to carry over.

How to Trade It (and Whether to Bother)

Scenario 1 — gap-producing markets (stocks, indices). If you trade session-based instruments and find a textbook Abandoned Baby, treat it like a strong morning/evening star: enter on the close of candle 3 or on a breakout from the pattern's high (low), stop beyond the doji — its extreme is a natural invalidation point since it's the island. The bullish variant has better numbers behind it than the bearish one; after a bearish one, don't expect a deep move (rank 64).

Scenario 2 — crypto: trade the equivalent, not the legend. On BTC/ETH, don't wait for a pattern that can't print. Its functional equivalent on a 24/7 market is the morning doji star / evening doji star — the same narrative (impulse → doji indecision → impulse the other way), just without gaps. It has its own, measured, weaker statistics — trade by those.

Scenario 3 — CME and weekend gaps. BTC futures on CME have sessions and real gaps — the technical definition can be satisfiable there. Remember, though, that it's a different instrument with a different price path than spot, and the historical sample of such setups is microscopic. A research curiosity, not a strategy pillar.

How NOT to play it. Don't build a strategy around a pattern that occurs once every few thousand candles — even if the statistics were rock-solid (they're not), frequency alone makes it useless as a systematic tool. Don't trust scanners that "find" an abandoned baby on crypto spot — check the gap definition before you believe it. And don't cite a rank of 1/103 without the footnote "n=14."

Stop loss and target. On session-based markets: stop beyond the doji's extreme, target at the nearest significant level — close for the bearish variant, since its moves are shallow. The gap between the doji and candle 3 can be wide, so calculate your risk-reward before entering, not after.

Myth vs. Measurement

Myth: "Abandoned Baby is one of the most effective patterns in the history of research." Measurement: 70% reversals and rank 9/103 is true — calculated from 293 cases out of 4.7 million candles, with key subsets holding only a dozen or so observations. Sample size is part of the statistic. Without it, a percentage is an anecdote.

Myth: "I found an abandoned baby on the BTC chart." Measurement (or really, the definition): without a full gap between the wicks, this pattern doesn't exist, and crypto spot doesn't create gaps. What you're looking at is a morning or evening doji star — a related setup with its own, more modest numbers.

Myth: "The bullish and bearish variants perform the same." Measurement: they reverse at similar rates (70% vs 69%), but performance is worlds apart — rank 9 vs 64. After a bearish Abandoned Baby, the move tends to be short and shallow. Symmetry of shape doesn't mean symmetry of results.

A Sample Scenario — and What to Look for Instead on BTC

Textbook version, equity market: a stock, after a string of declines, gaps down on the open, prints a doji hanging entirely below the wicks of its neighbors, and the next day gaps up and closes bullish. Enter on the close of the third candle, stop below the doji's low, target at the resistance along the way — with partial profit-taking and the awareness that you're trading a pattern whose historical sample is paper-thin.

Real version, BTC/USDT D1: after a week of declines, a red candle prints, then a doji, then a strong green candle — with no gaps at all, because there can't be any. Don't call this an abandoned baby. It's a morning doji star — and its statistics and its rules of engagement (confirmation on a close above the pattern's high, stop below the doji's low) are the ones that apply. The difference isn't academic: assigning an island pattern's statistics to a gap-free setup inflates your actual edge by several classes.

Quick checklist:

• Does the market even produce gaps (stocks/indices/futures — not crypto spot)?
• Is the doji separated by a gap from BOTH neighboring candles — wicks included?
• Is there a clear trend to reverse before the pattern?
• Do you remember the key statistics rest on a dozen to a few hundred cases?
• On crypto: are you trading the morning/evening doji star by its own numbers?
• Does a stop beyond the doji's extreme give a sensible risk-reward ratio?

Abandoned Baby is a beautiful pattern and an even better test of a careful statistics reader: rank 9/103 on the cover, n=293 in the footnote, and zero occurrences on the market most of us actually trade. Whoever reads only the cover hunts a unicorn. Whoever reads the footnote trades what's actually on their chart.

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