Fibonacci, Elliott & Harmonics
Harmonic Trading: Volume One
Carney's four patterns are not shapes you eyeball, they are arithmetic. A Gartley finishes at 0.786 of the first leg, a Bat at 0.886, a Butterfly at 1.27, a Crab at 1.618. Learn those four numbers and the B point that sorts them and you have the whole method. The remaining 250 pages are annotated charts of trades that worked.
The one idea
Every harmonic pattern is the same five dots on a chart, labeled X, A, B, C and D. Price runs from X to A, pulls back to B, bounces to C, then makes a final push to D. That is it. A five-point zigzag. Traders have drawn that shape for a century and called it an M or a W.
Carney's contribution is refusing to let the shape be the signal. He measures every leg as a percentage of another leg and says the percentages, not the outline, decide what you are looking at. Two charts that look identical are different trades if the B point sits at 61.8 percent of the XA leg instead of 50 percent. One is a Gartley that should turn at 78.6 percent of XA. The other is a Bat that should turn 10 percent deeper, at 88.6 percent. Buy at the Gartley level in a Bat structure and you are early by exactly that 10 percent, with a stop that gets taken out before the real turn.
The payoff is that a pattern stops being a picture and becomes a price. Once you know the shape is a Bat, you can compute the exact dollar level where it completes before price gets there, along with two or three other Fibonacci projections that should land in the same small band. Carney calls that band the Potential Reversal Zone. You get a specific entry, a specific invalidation just past the zone, and a specific first target, all calculated in advance from swings that have already happened. Whatever you think of the theory, that is a real operational upgrade over eyeballing support.
What it actually teaches
Every trade in the book runs through the same six steps, in this order.
FIND THE FIVE POINTS
Locate a completed swing X to A, a pullback to B, a bounce to C, and a fifth leg pushing toward D. Carney labels legs XA, AB, BC and CD rather than using Elliott's wave numbers. The structure must be visually distinct with roughly symmetrical legs. Nothing is measured until all four points X, A, B and C are already printed on the chart.
MEASURE B TO NAME THE PATTERN
The B point retracement of the XA leg is the sorting key, and it is the one thing Carney insists you check first. Exactly 0.618 makes it a Gartley. Anything under 0.618, ideally 0.50 or 0.382, makes it a Bat. Exactly 0.786 makes it a Butterfly. 0.618 or less, with 0.382 and 0.50 common, makes it a Crab. 0.886 or slightly deeper without breaking X makes it a Deep Crab. Get B wrong and every number downstream is wrong.
CHECK C AND PROJECT BC
C must retrace between 0.382 and 0.886 of the AB leg for every pattern in the book, which in practice means one of six allowed values: 0.382, 0.50, 0.618, 0.707, 0.786 or 0.886. The BC projection then extends that leg forward and must land in the range the pattern allows. Gartley uses 1.13 to 1.618 and never more than 1.618. Bat needs at least 1.618 and can run to 2.618. Butterfly wants 1.618, 2.0 or 2.24. Crab needs an extreme 2.618, 3.14 or 3.618. Carney's reciprocal rule is that the C retracement predicts the BC projection: 0.618 pairs with 1.618, 0.50 with 2.0, 0.786 with 1.27, 0.886 with 1.13.
CONFIRM THE AB=CD
Underneath every pattern sits an AB=CD, where the CD leg travels the same distance as the AB leg. An equivalent AB=CD is the minimum requirement for a valid setup in every pattern except the Crab, where Carney says it matters least. Bats and Butterflies usually run an alternate 1.27 AB=CD, meaning CD is 1.27 times AB. Crabs usually run a 1.618 alternate. A perfect AB=CD is a 0.618 C point with a 1.618 BC projection and legs of equal time duration.
STACK THE NUMBERS INTO THE PRZ
Project the pattern's completion level from the XA leg: 0.786 for a Gartley, 0.886 for a Bat, 1.27 for a Butterfly, 1.618 for a Crab or Deep Crab. Plot the BC projection and the AB=CD completion. Three or more of those numbers landing in a tight band is the Potential Reversal Zone and the only place you are allowed to trade. Carney gives no numeric tolerance anywhere in this edition for how close is close enough. The word tolerance does not appear in the book.
EXECUTE, THEN MANAGE BY THE NUMBERS
Enter on the first test of the zone. Stop goes just beyond the far edge of the PRZ, which Carney calls the Stop Loss Zone. First profit target is the 0.382 or 0.618 retracement of the pattern's own range, high to low, taken on part of the position. If price blows through 0.382 on a gap or an outsized bar, aim for 0.618 instead. After that, trail a 0.382 stop measured from the reversal point to the reversal extreme. The governing rule is never let a profit become a loss.
What it looks like Monday morning
You stop calling every five-point zigzag a Gartley. Before you draw anything else, you measure the B point against the XA leg. That one measurement tells you whether the turn should come 10 percent shallower or 10 percent deeper than you were about to buy, and on a two-point stop that difference is the whole trade. On charting platforms with a built-in harmonic tool this takes about fifteen seconds.
The other thing that changes is where the stop goes. Instead of picking a round number, you compute the completion levels first, see whether three of them cluster, and put the stop just past the widest of the cluster. If they do not cluster, there is no zone and no trade. That single filter kills more setups than anything else in the book and is the most useful discipline in it. The exit machinery, first target at 0.382 or 0.618 of the pattern range and a 0.382 trailer after that, works on any entry method you already use and does not require you to believe a word about Fibonacci.
| Pattern | B point (of XA) | C point (of AB) | BC projection | AB=CD | D point completion |
|---|---|---|---|---|---|
| Gartley | 0.618 exact | 0.382 to 0.886 | 1.13 to 1.618, never above | Equivalent AB=CD | 0.786 of XA |
| Bat | Under 0.618, prefer 0.50 or 0.382 | 0.382 to 0.886 | 1.618 minimum, up to 2.618 | Alternate 1.27 preferred | 0.886 of XA |
| Butterfly | 0.786 exact | 0.382 to 0.886 | 1.618 minimum, 2.0 or 2.24 common | Alternate 1.27 most common | 1.27 of XA, never 1.618 |
| Crab | 0.618 or less, 0.382 and 0.50 common | 0.382 to 0.886 | 2.618, 3.14 or 3.618 | Alternate 1.618 preferred | 1.618 of XA |
| Deep Crab | 0.886, may exceed but not break X | 0.382 to 0.886 | 2.24 minimum, up to 3.618 | Alternate 1.27 typical | 1.618 of XA |
Every pattern shares the identical C point band, so the D point and the B point are the only numbers doing real work.
Harmonic Trading frequently looks better in the past than it does in the present or the future.Scott M. Carney, Harmonic Trading: Volume One
Where it fails
- Only one number per pattern is actually fixed. The book sells precision, but read the requirement lists and the precision is one number deep. A Bat's D point must be 0.886 of XA. Everything else is a band: B is anything below 0.618, C is anywhere from 0.382 to 0.886, BC runs 1.618 to 2.618, and the AB=CD can be equivalent or 1.27. The Crab is the same, one fixed 1.618 D point wrapped in wide ranges. With six allowed C values and multiple allowed BC values, a five-point structure that fails one pattern's test usually satisfies another's. Carney writes that no variation is permitted from the prescribed alignments. The prescribed alignments are mostly ranges.
- There is no statistical validation of any kind in 262 pages. Search the full text and the words backtest, win rate, sample size and success rate do not appear once. Not a single pattern is given a hit rate, an average win, an average loss, or an expectancy. Carney calls the Bat probably the most accurate pattern in the entire Harmonic Trading arsenal and offers nothing behind it. He is candid about why: in the introduction he describes compiling hundreds of charts, keeping the relationships that repeated, and says he went with the techniques that worked. That is in-sample selection described in the author's own words, and no out-of-sample test follows anywhere in the book.
- Failed patterns get renamed instead of counted as losses. This is the sharpest evidence that the patterns are imposed rather than found. Carney writes that the Deep Crab pattern evolved from two patterns, invalid Bat patterns and a specific type of Crab pattern. A structure that violated the Bat's rules did not go in the loss column, it became a new pattern with a new completion level. A system that spawns a category every time reality misses the target can never be falsified, and it explains why the pattern count kept growing across his later books instead of the win rate being reported.
- The chapter summaries contradict the chapters they summarize. The Butterfly summary paragraph says the BC projection must be at least a 1.618 extension. The numbered list directly underneath it, on the same page, says the BC projection can vary from 1.13 to 3.618. That list also omits the 0.786 B point and the 1.27 XA projection entirely, the two requirements the same page calls the pattern's defining elements, and instead reproduces the AB=CD chapter's summary list almost word for word. The Gartley has the same problem: the chapter text allows a 1.13, 1.27, 1.41 or 1.618 BC projection and the Newmont Mining example trades a 1.41, but the summary list narrows it to 1.27 or 1.618. If you learn this method from the summary boxes, which is how most people will use the book, you will be trading rules the chapters do not support. This is the 2010 first printing.
- It assumes a live data feed and a chair you do not leave. Most of the worked examples are 5-minute and 15-minute charts on index futures and currency pairs. Carney describes the entry window as a few price bars and admits he has missed major S&P 500 intraday completions because he left his screens for five minutes to get a cup of coffee. That is the honest version of the requirement. The pattern tells you a price, not a time, so you have to be watching when price arrives. Anyone trading around a job, or working off end-of-day data, is limited to the daily and weekly setups, which appear a handful of times a year per instrument.
- Publication did not freeze the method, and the market caught up. The book announces Volume Two on its own second page and Carney went on to add more patterns and, in later work, a momentum confirmation layer built on RSI. Adding a confirmation filter is an admission that the price ratios alone were not sufficient, which is the opposite of what this volume claims. Meanwhile every retail charting platform now ships an automatic harmonic scanner that flags Gartleys, Bats, Butterflies and Crabs on every instrument and timeframe. In 2010 finding these by hand was work. Whatever edge came from the effort is gone, and the intraday index-futures examples that fill these chapters come mostly from 2003 and 2004 contracts, before those exact order books went algorithmic.
- The rules occupy about a dozen pages. The rest is chart plates. There are 185 figures. The Bat chapter runs 26 pages, of which the definition takes two and the summary one, with the remaining 23 given to eight annotated charts whose commentary repeats the same sentences about tight zones and sharp reversals. Chapter 10 and Chapter 12 do show failures, roughly eight of them, which is more honesty than most pattern books offer, but they sit against fifty-plus winning examples all selected after the fact. You can extract the entire identification method into one table and lose nothing.
Who it is for
Buy it if
Buy it if you already trade Fibonacci retracements by feel and want a fixed rulebook that tells you which level to use and when. The four completion numbers and the B point sorting rule are worth the price on their own, and the trade management chapter, with its 0.382 and 0.618 first targets and 0.382 trailing stop, bolts onto whatever entry method you already run. It is also the primary source for the vocabulary, so if you deal with harmonic traders or platform scanners, this is where the definitions come from.
Skip it if
Skip it if you need evidence before you risk money, because there is none in here, not one hit rate or backtest. Skip it if you cannot watch a screen intraday, since the entry window is a few bars wide and most examples are 5 and 15 minute charts. And skip it if you already own The Harmonic Trader or use a platform with a harmonic scanner, because you have the ratios already and the extra 250 pages are charts of trades that went the right way.
As an Amazon Associate I earn from qualifying purchases. It costs you nothing extra, and it does not change what gets recommended or what the verdict says.