Deep crab pattern — the deeper variant of Carney's crab

Scott Carney describes Deep Crab as a refinement involving invalid Bat formations and a specific Crab variant. Its deeper B reference differs from regular Crab, while D still references 1.618 XA. Check the remaining projections and the chosen definition together. The worked examples show confluence and anchor sensitivity, without treating a deep or narrow zone as evidence of a profitable reversal.

What the Deep Crab is and where it came from

The Deep Crab is a five-point X-A-B-C-D harmonic formation. Its B retracement is approximately 0.886 of X-A, and its D completion is near the 1.618 X-A extension measured from A. Do not confuse the Deep Crab's B ratio with the Bat's D ratio: the Bat normally has B below 0.618, preferably near 0.382 or 0.50, and D near a 0.886 X-A retracement.

The basics of trading harmonic patterns explain the measurements. Deep Crab is a defined variant, not necessarily a tighter or more profitable pattern. State the pivot-selection rule, availability time, ratios and tolerances before inspecting later outcomes.

Structure and Fibonacci levels

Deep Crab classification combines B near the 0.886 X-A retracement with its D extension and the remaining projection requirements.

For bullish geometry with X below A, measure AB/XA for B, BC/AB for C, AD/XA for D and CD/BC for the final projection. The author’s text allows B at 0.886 or deeper while remaining above X. A numerical tolerance cannot override that structural boundary, as the separate diagnostic diagram shows.

The bullish XA reference is D = A − 1.618(A − X), so XD = 0.618 XA. For the bearish mirror with X above A use D = A + 1.618(X − A). Carney’s current page names BC projections 2.24, 2.618, 3.14 and 3.618; the cached Volume One text instead discusses a BC minimum of 2.0 and variations up to 3.618. Record which version and named levels you use rather than silently combine them into one continuous band. Deep Crab also gives equivalent or alternate AB=CD projections greater importance. Check the formulas with Fibonacci extensions and Fibonacci retracements; shared anchors do not create independent predictive confirmations.

Being between 2.24 and 3.618 does not itself match a named BC coefficient. The original candidate’s nearest listed value is 2.24, about 0.097805 away; a separately chosen ±0.02 coefficient tolerance would reject it. That tolerance is our test assumption, not Carney’s universal rule. Nor do its final legs match a 1.27 AB projection closely. Do not call a candidate complete by discarding failed chosen checks.

A second deliberately constructed model makes confluence explicit: X = 1.12000, A = 1.15000, B = 1.12280 and C = 1.13599. AB is 272 pips and BC is 131.9 pips, giving AB/XA ≈ 0.906667 and BC/AB ≈ 0.484926. Exact B = 0.886 XA would be 1.12342; our B is 6.2 pips deeper and still above X.

At the XA D, CD/BC ≈ 2.617892 and CD/AB ≈ 1.269485. Highest minus lowest projection is 0.14 pips. This manufactured geometric confluence does not establish a 0.14-pip stop, future price accuracy or how often such structures occur. Round only the displayed prices; permitted order prices depend on the instrument tick size.

How to recognise a Deep Crab step by step

Step 1 — find the X-A opening leg

Record the linear scale, ordered anchors and their confirmation times. A visually sharper opening leg does not establish more reliable predictions. Freeze the pivot rule so later price information cannot move X or A within an earlier test result.

Step 2 — check that point B reaches 0.886

Check whether B is near the 0.886 retracement of X-A. A shallower B does not automatically identify a classic Crab: the other points and their ratios must also meet that pattern's specification. The classic Crab uses B in the approximate 0.382–0.618 range.

Step 3 — measure the B-C leg

Calculate BC/AB explicitly and record the C filter required by your selected implementation. In the constructed model C is between B and A, with ratio about 0.484926. Do not infer that every candidate satisfying this one location or ratio qualifies without B, D, BC projections and AB=CD checks.

Step 4 — confirm point D with a double measurement

Compare each selected BC and AB=CD projection with the XA D reference under a tolerance defined in advance. Distinguish the envelope of all projections from the intersection of their permitted ranges. Projected D, a later-confirmed turning extreme, a trade signal and an actual fill are separate records.

Entry, stop and targets — a hypothetical example

Define the entry, cancellation, exits and fill assumptions separately. In an independent fictional USD-account model, buy EUR 10,000 at Ask 1.10200 and assume an actual stop sale fill at Bid 1.09900 or target sale fill at Bid 1.10800. Price outcomes are −USD 30 or +USD 60. With USD 4 total additional opening/closing costs, net outcomes are −USD 34 or +USD 56. Spread is already represented by the assumed Ask/Bid fills and must not be subtracted twice. An ordinary stop may execute worse: a sale at 1.09850 gives −USD 39 net.

If choosing additional target rules at 0.382 and 0.618 AD from D, the constructed model’s AD = 485.4 pips gives references 1.12000228 and 1.13145772. They are not automatically expected exits. A separately chosen 1.732 XA buffer would be 1.09804, not a universal Deep-Crab stop requirement. Actual entry, exit sizes, quote-side fills and costs determine reward:risk. No observed frequency of overshooting or precise entry is established here.

Keeping A = 1.15000 and raising X from 1.12000 to 1.12100 reduces XA to 0.02900 and raises D to 1.103078, a 16.18-pip change. Fixed formulas are reproducible, but a changed anchor requires rechecking every condition. A separate B-sensitivity exercise holds X, A and C = 1.13599 fixed: BC D = C − 2.618(C − B) and AB D = C − 1.27(A − B).

Raising B by 10 pips moves BC D up 26.18 pips and AB D up 12.70 pips, while XA D stays unchanged. The spread is highest minus lowest of all three references, not just BC minus AB. These rows isolate anchor sensitivity; they do not all qualify as Deep Crab. In particular the last B has AB/XA ≈ 0.873333, below the chosen minimum 0.886. Narrow confluence in the middle row is not robust to every alternate anchor and gives no probability of reversal.

Deep Crab versus the classic Crab and related patterns

The classic Crab uses a shallower B, approximately 0.382–0.618 of X-A, while the Deep Crab uses B near 0.886. Both have a D reference near the 1.618 X-A extension, but their other projection requirements also matter. The Bat instead places D near the 0.886 X-A retracement and normally has B below 0.618. Classification requires the complete structure, not an isolated ratio.

Who this pattern is for

Compare complete definitions with the Gartley pattern, Bat pattern and broader technical analysis. Deep B alone does not establish a tighter zone or a trading advantage. Retain rejected and failed candidates, freeze the rules and assess later data with costs instead of altering anchors to fit a known reversal.

What to do tomorrow to master the Deep Crab

1. Choose a series, interval and pivot rule, then record when X, A, B and C become known. Check B depth and the separate X boundary.
2. Record each ratio’s denominator, selected BC levels, AB=CD version and tolerance. Retain failures of individual checks.
3. Use an alert to locate a candidate for review; do not equate it with confirmation, an available fill or a complete geometric test.
4. Evaluate frozen entry/exit rules on later data with size, quote sides, costs and missed fills. Twenty demo trades or a repeatable hit rate alone does not establish future profit or live readiness.