Harmonic Patterns — Gartley 222, Butterfly and Bat on XABCD Points

On the twenty-third of January 2024, on the daily GBP/USD chart, Agnieszka — a trader with a decade of experience on the Warsaw FX desk — drew five points on her screen. X sat at 1.2670 from the December high, A at 1.2515 from the low of that same month, B at 1.2611, C at 1.2530, and finally a projected D at 1.2552. The ratios between the legs matched, to three decimal places, the classic Bat pattern Scott Carney had codified back in 2001: AB measured 49.3 percent of the XA retracement, and point D was projected to the 88.6 percent retracement. Five days later, when the daily candle closed at 1.2549 with a long lower wick and a bullish engulfing showed up on the four-hour chart, Agnieszka opened a long position. Within three weeks price reached 1.2680, leaving 1,310 pounds of profit on a single standard lot. This article explains why a geometry built on Fibonacci retracements has become one of the most precise methods of trading the Forex market and what truly separates the Gartley 222, the Butterfly and the Bat.

Harmonics — the geometry of price according to H.M. Gartley

Harmonic patterns trace their origin to an unusual book published in 1935 in New York by the Lambert-Gann house. Harold McKinley Gartley, a Wall Street advisor and analyst, released "Profits in the Stock Market" — the first complete treatise to combine Fibonacci cycles with the recognition of recurring price formations. On page 222 he sketched the structure that the entire technical-analysis community would later come to call, quite simply, the "Gartley 222". Decades on, Larry Pesavento, Bryce Gilmore and finally Scott Carney would add new members to that family — Butterfly, Bat, Crab and Cypher — but it is Gartley's original idea that remains the foundation of them all.

The idea itself is simple, and at the same time deeply surprising. Price action on every chart is made up of impulses and corrections, and each correction tends to halt around one of a handful of specific levels drawn from the Fibonacci sequence: 38.2 percent, 50 percent, 61.8 percent, 78.6 percent. Gartley noticed that certain combinations of these retracements, arranged into a five-point XABCD structure, made it possible to anticipate a reversal in price ahead of time.

Gartley 222 — the pattern of patterns, still alive after ninety years

The classic Gartley 222, despite nine decades on the clock since its publication, remains the formation most harmonic traders start with. The reason is pragmatic: its ratios are relatively forgiving and it appears on charts several times more often than the more exotic Crab or Cypher. Leg XA is the initial impulse, AB a retracement of 61.8 percent, BC a correction between 38.2 and 88.6 percent of AB, and CD the third leg whose point D lies at the 78.6 percent retracement of XA. In a classic Gartley point D never extends beyond the X-to-A range — the pattern is contained inside the original move, which sets it apart from the later Butterfly and Crab constructions.

The bullish variant starts at a high (X), descends to a low (A), bounces 61.8 percent back up (B), pulls back down (C) and drops again to touch the 78.6 percent retracement of leg XA — that final low is point D, where you open the long position. The bearish variant is its exact mirror image. Statistically the Gartley 222 produces a win rate of 58-62 percent when ratios are respected strictly — worse than the Bat, but still enough to deliver a solid edge with good position management.

Butterfly — the extension pattern, when price overshoots X

The Butterfly is sometimes described as "the extended Gartley". Larry Pesavento popularised it in the nineties through "Fibonacci Ratios with Pattern Recognition". What separates the Butterfly from the original Gartley is the depth of two of its legs. Point B reaches as deep as 78.6 percent of the XA retracement — considerably deeper than the classic 61.8 percent in Gartley. And point D, instead of halting inside the XA range, extends beyond X and lands at the 127 percent or, in the aggressive variant, the 161.8 percent extension of leg XA.

The mechanical reading of a Butterfly is this: price breaks the original high or low (X) by 27 to 62 percent, generating what is known as a false breakout with exhaustion. Traders late to the trend climb on board at precisely point D, providing liquidity to those who understand the geometry. Stop loss is placed 5 to 15 pips beyond point D on H4 or 20 to 40 pips on D1. The average win rate on the Butterfly sits at 55-60 percent, but because point D extends beyond X the reward-to-risk usually lands at 1:3 to 1:5 — the highest in the entire harmonic family.

Bat — the most rigorous pattern of the Carney era

The Bat was developed by Scott Carney in 2001 and described in his first book, "The Harmonic Trader", in 2004. Carney was not happy with how loosely traders treated the proportions of the original Gartley — many of them accepted a B retracement anywhere between 50 and 78.6 percent as "close enough to 61.8". In his view, that kind of flexibility completely broke the statistical edge of the pattern, so he engineered a new formation with much tighter requirements. In the Bat, point B sits at the 38.2 or 50 percent retracement of XA — shallower than in the Gartley — and point D reaches as deep as the 88.6 percent retracement of XA, almost touching the original point X but never breaking through it.

The formation ends with a deep retest of the extreme, at which exhausted counter-trend speculators are forced to capitulate. For the trader, this translates into an entry much closer to point X than in a Gartley — which shortens the stop loss and lifts the win rate, because the 88.6 percent level has historically acted as strong support on most currency pairs. Carney and independent researchers report a Bat win rate of 63-67 percent on the daily chart — the best of all the classic harmonics. Tolerance is plus or minus 2 to 3 percent here, not the 10 that traders sometimes allow themselves in "flexible" Gartleys.

The PRZ zone and candle signal — where you actually enter the market

Point D is not a single price value, but a zone where several Fibonacci levels converge. In the literature it is called the PRZ — Potential Reversal Zone. In the ideal case, three values meet inside the PRZ: the XA retracement (78.6 or 88.6 percent depending on the pattern), the BC extension (127 or 161.8 percent) and the alternate AB=CD projection. The tighter the convergence zone (ideally 10 to 25 pips on D1, 5 to 12 pips on H4), the higher the probability of a reaction.

Entering on the simple touch of the calculated D level is a beginner's mistake. Experienced harmonic traders wait for candle confirmation inside the PRZ. The most common signals are a bullish or bearish pin bar, an engulfing pattern in the opposing direction, a double bottom or double top on a lower timeframe, or a classic morning star or evening star. Without that confirmation, harmonics lose part of their statistical edge and drop to around 50 percent — coin-flip territory.

"Harmonic patterns are not mechanical money-making machines. They are a geometric map showing where, with the highest probability, a reversal will take place. Whether you actually make money out of them depends on the discipline of respecting the ratios, the patience to wait for candle confirmation inside the PRZ, and the consistency of how you manage the position." — Scott M. Carney, "Harmonic Trading, Volume One", FT Press, 2010, from the chapter on the psychology of harmonic patterns.

Case study — the Bat on GBP/USD in January 2024

Agnieszka's decision was not taken in isolation. Point D in this Bat fell exactly on a previous support level from September 2023 — which added a structural confluence independent of the Fibonacci mathematics. Her trading journal showed that Bat configurations with an additional structural support carried a historical win rate of 71 percent in her data, compared with a 65 percent baseline for the Bat alone. That five-point difference, at the scale she trades, translates into tens of thousands of pounds of additional profit annually.

The most common mistakes in trading harmonic patterns

Harmonics look easy — five points, a handful of retracements, done. In practice most traders entering the world of harmonic patterns make the same mistakes, the kind that turn signals with a 65 percent edge into a losing run of trades.

• Stretching the ratios. When point B sits at the 67 percent retracement of XA rather than at 61.8, the beginner tells themselves: "close enough". It is not. Carney defines tolerance at 2 to 3 percent in either direction. Anything outside that band is no longer a pattern; it is a random configuration of prices.
• No candle confirmation inside the PRZ. The single most common source of losses. Price touches the computed point D, the trader enters immediately, price punches another twenty pips through the PRZ and takes out the stop loss. The patience to wait for an engulfing or a pin bar inside the zone is the difference between 50 percent and 65 percent in win rate.
• Trading harmonics against the higher-timeframe trend. A bullish Bat inside a strong daily downtrend is a textbook counter-trend trap. Statistically such configurations drop to roughly 45 percent in win rate despite a perfect geometry.
• Closing the position too early. The three-step target system Carney designed is an integral part of the strategy. A trader who books the whole position at TP1 leaves, on average, 1.5 points of reward-to-risk on the table.
• The wrong timeframe. Harmonics were designed for daily and weekly charts. On M5 or M15 they generate so much noise that any statistical edge dissolves. The sweet spot is D1 for swing positions and H4 for intraday work.

A practice plan — from your first Gartley to proficiency on the Bat

Mastering harmonic patterns takes between three and six months of systematic chart work. Below is a study schedule, road-tested by hundreds of traders sharing their experience on the HarmonicTrader.com forums.

1. Month one — learning to spot the Gartley 222. Pick one currency pair (EUR/USD or GBP/USD) and the daily timeframe. Scroll the chart back three years. Using TradingView's XABCD tool, locate every potential Gartley. You should find between twelve and twenty. Check in how many of those cases price actually reversed from point D.
2. Month two — add Butterfly and Bat. Repeat the same process for the next two patterns. Butterflies appear about half as often as Gartleys, and Bats roughly with the same frequency as Butterflies. Compile your own win-rate statistics for each of the three.
3. Month three — paper trading on a demo account. Trade only those patterns that meet three criteria: strict ratios, confluence with structural levels, and a candle confirmation inside the PRZ. The target is at least twenty trades. After that sample size you have enough data to judge whether your actual win rate is in line with the expected 60-65 percent.
4. Months four to six — live account with minimal risk. One percent of capital per trade, no more. The first fifty trades go through without any strategy modifications. After fifty positions, analyse your journal and adjust — for example skipping Gartleys in counter-trend conditions if your data shows that is where most of your losses come from.

Harmonic patterns, first described by H.M. Gartley in 1935 and refined by Scott Carney in the first decade of the twenty-first century, are geometric XABCD formations built on strict Fibonacci retracements. The three most important members of the family differ in the depth of the AB and CD legs: Gartley 222 stays inside the XA range (D at 78.6 percent), Butterfly extends beyond X to the 127-161.8 percent extension, and Bat retests the extreme at the 88.6 percent retracement of XA. The key to making them work lies in three things: respecting the ratios strictly with a tolerance of 2 to 3 percent, waiting patiently for candle confirmation inside the PRZ, and applying the three-step profit-target framework. With those pillars in place harmonics deliver a statistical 60-67 percent win rate at an average reward-to-risk of 1:2.5 to 1:3.2 — on D1, W1 and H4. Without them, you are left with a random trade with roughly a 50 percent hit rate.

Related reading: Fibonacci retracements — the basics for the mathematics behind harmonics; Fibonacci extensions — setting targets as a companion piece to the BC projection topic; Elliott wave theory — the historical context of market geometry and its links to Gartley's framework.