Harmonic XABCD Patterns and Elliott: Check the Measurements
Worked Gartley, Butterfly, Bat and Crab measurements with declared conventions, original data and a separate confirmation protocol.
Published: (UTC+3)
Русский · English · Deutsch · Español · Français · Italiano · 繁體中文 · 日本語 · 한국어
Harmonic geometry and Elliott structure examine different requirements. These invented data sets make the measurements reproducible without treating a drawn pattern as a trading result.
Define the vertices and the ratio convention
For a rising XA followed by a downward completion, use positive lengths XA=|A−X|, AB=|B−A|, BC=|C−B| and CD=|D−C|. The terminal ratio here is AD/XA, not XD/XA. XABCD gives five vertices and four legs; it does not establish a five-wave Elliott impulse.
These examples select published variants explicitly. Our comparison tolerance ε=0.02 is an absolute ratio difference chosen for the exercise, not a universal software setting. The source pages contain differing projection variants. Read the actual ratios in the tables; a source name or an attractive outline is insufficient.
| Measure | Calculated value |
|---|---|
| XA | |A−X| |
| AB | |B−A| |
| BC | |C−B| |
| CD | |D−C| |
| AD/XA | |D−A| / |A−X| |
| ε | 0.02 |
Do five XABCD vertices establish five Elliott waves?
No. They mark four geometric legs. Elliott roles, degree and subdivisions require separate checks.
Gartley: a declared range-based example
X=100, A=200, B=138.2, C=183.2, D=121.4 give AB/XA=0.618, AD/XA=0.786 and equal AB/CD lengths. CD/BC is 1.373333. We use the continuous 1.13–1.618 BC range in Carney’s 2020 checklist; this is not the dedicated page’s narrower highlighted 1.27/1.618 alignment.
All prices are invented on a linear scale. Compare the selected version and every required relationship before considering the shape complete. A matching reference at D supplies no later reaction, order fill or Elliott validation.
| Measure | Calculated value |
|---|---|
| X (t=0) | 100 |
| A (t=10) | 200 |
| B (t=20) | 138.2 |
| C (t=30) | 183.2 |
| D (t=40) | 121.4 |
| XA | 100 |
| AB | 61.8 |
| BC | 45 |
| CD | 61.8 |
| AD | 78.6 |
| AB/XA | 0.618 |
| BC/AB | 0.728155 |
| CD/BC | 1.373333 |
| AD/XA | 0.786 |
| CD/AB | 1 |
| C−AB | 121.4 = D |
Does D=121.4 alone confirm an Elliott impulse?
No. It supplies a chosen harmonic endpoint relationship, not the impulse’s internal structures or later market behaviour.
Butterfly: an extension beyond X
X=100, A=200, B=121.4, C=151.6, D=73 give B’s 0.786 retracement and AD/XA=1.27, with AB=CD. CD/BC=2.602649 is within the chosen 0.02 of 2.618. The public source discusses several BC variants; this illustration selects that branch, not a universal formula.
All prices are invented on a linear scale. Compare the selected version and every required relationship before considering the shape complete. A matching reference at D supplies no later reaction, order fill or Elliott validation.
| Measure | Calculated value |
|---|---|
| X (t=0) | 100 |
| A (t=10) | 200 |
| B (t=20) | 121.4 |
| C (t=30) | 151.6 |
| D (t=40) | 73 |
| XA | 100 |
| AB | 78.6 |
| BC | 30.2 |
| CD | 78.6 |
| AD | 127 |
| AB/XA | 0.786 |
| BC/AB | 0.384224 |
| CD/BC | 2.602649 |
| AD/XA | 1.27 |
| CD/AB | 1 |
Is D=73 below X=100 automatically a drawing error?
Not in this selected extension construction. Its terminal convention differs from the retracement used for Gartley.
Bat: distinguish the B retracement
X=100, A=200, B=150, C=174.9, D=111.4 give AB/XA=0.5 and AD/XA=0.886. CD/BC=2.550201 lies in the selected 1.618–2.618 range, while CD/AB=1.27. The B retracement separates this example from our 0.618 Gartley case.
All prices are invented on a linear scale. Compare the selected version and every required relationship before considering the shape complete. A matching reference at D supplies no later reaction, order fill or Elliott validation.
| Measure | Calculated value |
|---|---|
| X (t=0) | 100 |
| A (t=10) | 200 |
| B (t=20) | 150 |
| C (t=30) | 174.9 |
| D (t=40) | 111.4 |
| XA | 100 |
| AB | 50 |
| BC | 24.9 |
| CD | 63.5 |
| AD | 88.6 |
| AB/XA | 0.5 |
| BC/AB | 0.498 |
| CD/BC | 2.550201 |
| AD/XA | 0.886 |
| CD/AB | 1.27 |
Can the 0.5 B ratio simply be relabeled as our Gartley?
No. The selected Gartley uses 0.618 at B. Check the other Bat measurements too; one ratio does not confirm the pattern.
Crab: main convergence and a secondary reference
X=100, A=200, B=138.2, C=176.4, D=38.2 give regular B with AB/XA=0.618, AD/XA=1.618 and CD/BC=3.617801, near 3.618. CD/AB=2.236246. The secondary reference C−1.618AB=76.4076 does not meet D. We do not turn that descriptive reference into a universal requirement or an automatic invalidation of Crab. The public page also mixes in a Deep Crab description, which this example does not adopt.
All prices are invented on a linear scale. Compare the selected version and every required relationship before considering the shape complete. A matching reference at D supplies no later reaction, order fill or Elliott validation.
| Measure | Calculated value |
|---|---|
| X (t=0) | 100 |
| A (t=10) | 200 |
| B (t=20) | 138.2 |
| C (t=30) | 176.4 |
| D (t=40) | 38.2 |
| XA | 100 |
| AB | 61.8 |
| BC | 38.2 |
| CD | 138.2 |
| AD | 161.8 |
| AB/XA | 0.618 |
| BC/AB | 0.618123 |
| CD/BC | 3.617801 |
| AD/XA | 1.618 |
| CD/AB | 2.236246 |
| C−1.618AB | 76.4076 ≠ D |
Do all secondary references mentioned by the author converge?
No. XA and BC converge here, but the secondary 1.618AB projection does not end at D. Preserve that observation without treating this one reference as a universal requirement.
A potential zone and a reaction are different observations
A PRZ is a candidate area derived from declared measurements. In our invented Gartley case D=121.4 at index 40, add completed closes 122.4 at 41 and 123.4 at 42. They satisfy an authored observation rule: two later completed closes above D. This rule is not represented as Carney’s official entry method.
The two closes do not establish all intrabar extremes, a lasting reversal, an executed order or a valid Elliott count. The diagram stops at 42; a five-bar review ending 45 is still incomplete. Keep geometry, observed reaction and the eventual research outcome separate.
| Measure | Calculated value |
|---|---|
| D (t=40) | 121.4 |
| t=41 | 122.4 > D |
| t=42 | 123.4 > D |
| t=45 | — |
| D−2 | 119.4 |
Can two closes above D establish the whole five-bar outcome?
No. Later bars and execution are missing. Record the observed reaction and keep the remaining outcome unresolved.
Research the rule before calling it a strategy
Before testing new cases, freeze pivot selection, the source, interval, ratio version, tolerance and the event to observe. For this exercise choose two completed closes>D within five later bars, and a separate failure condition completed close<D−2. For D=121.4 that boundary is 119.4.
Choose event ordering, cancellation, fills, costs and all eligible cases before viewing outcomes. A newly selected pivot or a changed tolerance creates a different case. These hand-built examples test arithmetic and definitions; they report no empirical hit rate or profitable strategy.
| Measure | Calculated value |
|---|---|
| D | 121.4 |
| D−2 | 119.4 |
| t(D)+5 | 45 |
May the tolerance be widened only after a losing outcome?
That changes the protocol after the fact. Preserve the original test and report the new version separately.