ABCD pattern — a trading guide to the harmonic building block

ABCD labels four successive turning points joined by three price legs. The measurements are useful only when you state the variant, each ratio’s denominator and how the points become known. This guide separates equal AB=CD from alternate projections and from the additional rules needed to test a trade.

What the ABCD pattern is and why it is foundational

In Scott Carney’s AB=CD definition, BC is a move against AB and CD has a similar price length to AB. His method uses BC retracements of 0.382–0.886 and a complementary BC projection. This is a classification of geometry, not evidence that the market must repeat the shape or reverse at its completion.

Related harmonic structures add points and specific ratios; this does not make every four-point zigzag a qualifying signal. The basics of trading harmonic patterns introduce that vocabulary. Keep the measurement rule separate from the trade’s outcome.

Three legs and the Fibonacci levels

Write positive lengths L = |B − A|, M = |C − B| and N = |D − C|. Then r = M/L measures BC/AB, q = N/M measures CD/BC and k = N/L measures CD/AB. By arithmetic, k = r × q. For equal legs, k = 1 and q = 1/r: deeper BC requires a smaller BC multiplier. A label such as “1.27” is incomplete without its reference leg. Drawing a level with a technical analysis tool does not turn it into a confirmed reversal.

For an equal-leg AB=CD, let r = BC/AB. The matching extension is CD/BC = 1/r: a deeper BC retracement requires a smaller BC extension to preserve CD = AB. For example, 0.618 pairs with about 1.618, while 0.786 pairs with about 1.272. A broader ABCD variant must state its own ratios rather than assume equal legs.

The last column rounds reciprocals to four decimals. Using 1.272 times this BC instead gives 157.2192 pips and D 1.07663808 before tick rounding. CD would be 78.6096% of AB, not equal to it. Popular rounded ratio pairs are approximate: 0.618 × 1.618 = 0.999924. Do not demand exact equality while using rounded labels as exact inputs.

The three variants of the ABCD pattern

Name the rule set rather than treating “classic”, “perfect” and “alternate” as universal trading grades. Price equality, a selected retracement, closeness to rounded projections and a time filter are separate conditions. Their coexistence does not establish rarity or superior reliability. State which conditions your test actually requires.

Carney’s alternate AB=CD projections scale AB by 1.27 or 1.618 from C, alongside other measurements. For the separate example above, 1.27 × 200 = 254 pips gives D 1.06696, while 1.618 × 200 = 323.6 pips gives D 1.06000. These are independent alternatives, not successive excuses for keeping a failed position open. Scaling AB is different from scaling the 123.6-pip BC.

Symmetry of time and price

Price-length equality is different from elapsed-time similarity. If a rule also compares candle counts, specify its threshold and test it separately; three or forty candles do not establish a universally weaker signal. Absolute and relative price tolerances also select different sets. In this hypothetical table, CD is always two pips longer than AB.

Both boundaries are inclusive. A 1% threshold permits only 0.4 pip for AB 40, but 4 pips for AB 400. Do not switch tolerance definitions after seeing the outcome. Preserve full precision in classification, then round display values; record the raw or tick-rounded price basis. These thresholds are our assumptions, not a universal Carney rule.

How to trade the ABCD pattern — entry, stop and targets

A projected D is a level to evaluate, not proof of an executed trade. Specify when your rule can confirm A, B and C: a swing requiring later candles was not known at its original timestamp. Choose entry and invalidation conditions before viewing the subsequent result. A reversal candle can be one test condition, but its name alone does not establish profitability or require every reader to buy the level.

A rule may choose 0.382 or 0.618 retracements of CD as targets, but geometry does not guarantee either fill or net profit. For a separate hypothetical USD-account long, assume full entry at Ask 1.07256, a pip of 0.0001, EUR 10,000 exposure, USD 3 total commission and no funding or conversion. An ordinary stop at 1.07056 would give USD 20 price loss and USD 23 net loss if filled there. Assume full Bid exits at the chosen target levels. Targets measured from D 1.07236 are 1.08000 and 1.08472: price gains USD 74.40 and USD 121.60, or USD 71.40 and USD 118.60 after commission. Spread is already inside the actual entry and exit-side prices. Net reward:risk ratios are about 3.10:1 and 5.16:1 under these assumptions; a worse stop fill changes the loss. This does not assign probabilities or simulate interim close-out. The basics of risk management develop the separate sizing decision.

The most common mistakes when trading the ABCD

1. Record which leg each ratio scales; a mismatch with your selected definition does not establish a trend change.
2. Separate the extreme’s timestamp from the time your rule confirms it.
3. Keep time and price tolerances distinct and fixed before reviewing outcomes.
4. Check entry/exit quote sides and stress an ordinary stop’s worse fill rather than assume a wider stop prevents loss.
5. Do not move points or rename a failed variant after seeing the result.

From the ABCD to more complex patterns

Five-point structures require their own measurements. Carney’s Gartley definition combines specified B and D retracements with AB=CD confluence; its D measurement is 0.786 of XA. The Butterfly pattern has separate rules. Adding an X point alone does not classify the structure or prove its performance. Use the linked lessons to compare definitions, while keeping this guide’s three denominators explicit.

What to do tomorrow to master the ABCD pattern

1. Use a fixed chart/data specification and select A, B and C with a rule known at the time. Mark their confirmation times separately from the original extremes.
2. Calculate BC/AB, CD/BC and CD/AB, and state whether equality or an alternate projection is required. Keep price precision and tolerance consistent.
3. If you add a time filter, define it before the test. Similar candle counts are a condition to evaluate, not proof of a cleaner signal.
4. Record the variant, available-time points, entry/exit sides, costs and actual result. Compare separately defined gross and net reward:risk rather than treating one ratio as a complete strategy.
5. Review examples outside the period used to choose the rules. Twenty demo trades or an apparent success rate does not prove live readiness; costs, execution and data limitations need their own checks.