The ABCD pattern: a description, not a forecast
ABCD labels a price sequence after it forms. Labelling is subjective, so fix the rules first, count completions and failures alike, and treat the measured-move target as a claim to test.
Start in plain English
The ABCD pattern names four points on a chart: a low (A), a rise to a high (B), a pullback to a higher low (C), and a second rise to a projected point (D). Many trading guides add a rule that the second leg tends to match the first, so D is projected as C plus the length of AB: the "measured move". For a small-cap scalper the pattern is attractive because it turns a messy chart into four clear levels: an entry near C, a stop under C and a target at D.
The core idea of this lesson: ABCD is a label applied to a sequence, and the label is easiest to apply after the sequence has finished. Before it finishes, different traders can label the same chart differently, and the projected D is a claim, not a destination. So the work is the same as for any setup: write labelling rules that two people would apply the same way, take the decision at a fixed point, and count the sequences that extended to D alongside those that broke below C.
This lesson's labelling rules:
- A: the lowest low in the 15 minutes before B.
- B: the highest high after A, confirmed once price pulls back at least 30% of the A-to-B distance.
- C: the lowest low of that pullback, valid only if it retraces 30% to 70% of AB, holds above A, and lasts at least three 1-minute bars. C is confirmed when a 1-minute bar closes above the previous bar's high.
- D (projection): C + (B − A).
These are choices, not standards. Many published versions use Fibonacci ratios instead (for example C near 61.8% of AB); pick one set and keep it fixed.
Two paths from the same start
All prices, sizes and costs in this example are Illustrative. Pellucid Grid is fictional.
The setup at 09:58 ET. The low of day was $2.31 at 09:34. Price then traded $2.40 at 09:42 (A under this lesson's rules), rose to $2.90 at 09:51 (B; AB = $0.50), and pulled back for six bars to $2.62 at 09:57, a 56% retracement. The 09:58 bar closed at $2.70, above the 09:57 high of $2.68, which confirms C at $2.62. Projected D = $2.62 + $0.50 = $3.12.
The plan, written before 09:58. Entry with a marketable limit order (ceiling $2.73), filled at $2.71. Stop at $2.60, two cents under C. Target $3.12, the projected D. Time stop 10:20. Planned risk $0.11 per share. Size: $50 risk ÷ $0.11 = 454 shares. Costs: $0.005 per share round trip (Illustrative), plus any slippage.
Path AThe pattern completes and extends
Price clears B at 10:04, trades $3.00 by 10:09 and $3.14 at 10:14. The $3.12 limit sell fills. A reviewer looking at the chart afterwards sees a textbook ABCD.
Path BThe pattern fails at C
10:00 trades back to $2.64; 10:01 breaks $2.62 and trades $2.57. The stop at $2.60 fills at $2.58. A reviewer looking at this chart afterwards sees no ABCD at all, only a lower low.
Path CA stall below B
Price rises to $2.86, short of B, and drifts between $2.70 and $2.84. At 10:20 the time stop exits at $2.74.
The review note for all three: "Plan followed; points labelled by the written rules; risk 1R; outcome A, B or C." Note what the two reviewers in Path A and Path B saw: one chart that "worked" and one that "was not an ABCD at all". That is the bias this lesson is about. The rest covers why labels disagree, what the measured-move claim does and does not say, and how to count without letting outcomes choose the sample.
Why labels disagree
Even with written rules, reviewers can disagree at each point. On the Pellucid Grid chart:
| Point | This lesson's label | A defensible alternative | Effect |
|---|---|---|---|
| A | $2.40 at 09:42 (lowest low in the 15 minutes before B) | $2.31, the low of day at 09:34 | AB becomes $0.59; the retracement 47%; D moves to $3.21 |
| B | $2.90, the 09:51 high | $2.85, the highest close (09:50) | AB and D shift by 5 cents; the retracement changes |
| C | $2.62, confirmed at 09:58 | A trader using a 61.8% rule waits for about $2.59 | No entry at all on this chart |
None of these alternatives is wrong. Each produces a different count. The danger is switching between them after seeing the outcome: labelling A at the low of day when that makes D land on the actual high, and at the local low when that fits better. A sample built that way describes the reviewer's choices, not the pattern.
Two practical checks reduce this:
- Label before revealing. Hide everything after the decision time, label, then reveal. The replay exercises in this lesson do exactly that.
- Label twice. Have a second person label the same charts with the same rules, or label them yourself again a week later. Count how often the labels agree. If agreement is low, the rules are not yet countable.
The measured-move claim
"CD equals AB" is the central claim of most ABCD material, sometimes with Fibonacci extensions (CD = 1.272 or 1.618 × AB). It is a description of some completed sequences, not a property of price. Three reasons it is hard to test:
- Selection. Sequences that reached D are remembered as ABCDs; sequences that failed at C are often not labelled at all, as in Path B.
- Flexible tolerance. "Close to D" without a stated tolerance can absorb almost any outcome.
- Many ratios. If 1.0, 1.272 and 1.618 all count, one of them is usually near wherever price stopped.
A fair test fixes the ratio and the tolerance before looking, and counts every labelled C, including those that broke down.
Observe vs. infer
At 09:59:00 in the Pellucid Grid example (fictional; figures Illustrative):
| What the screen shows (observed) | What a trader might infer | Why it is not established |
|---|---|---|
| The pullback from $2.90 to $2.62 retraced 56% of AB | "A healthy pullback" | Retracement depth is a measurement; "healthy" is a forecast |
| Volume fell from 310,000 at 09:51 to 109,000 at 09:57 | "Sellers dried up" | Lower volume fits fewer sellers and fewer buyers alike |
| The 09:58 bar closed above the 09:57 high | "C is the low" | C is confirmed by the rule; whether price stays above it is unknown |
| Projected D is $3.12 | "Price is headed to $3.12" | D is arithmetic on past points, not a destination |
Odds box
A distant target changes the shape of the results. Under these assumptions the plan breaks even at a Path A share of about 15%, so most trades end as Path B or C and long runs without a Path A are normal; the sample-size section puts numbers on that. No FennEdge data study currently covers ABCD; one is proposed above, including a labelling-agreement measure.
What shifts the odds
There is no FennEdge sample linking conditions to ABCD outcomes. These are hypotheses to record, not findings:
- Retracement depth of C (30%, 50%, 70% of AB).
- Time in the pullback (3 bars vs. 15 bars).
- Volume in the CD leg's first bars relative to the AB leg.
- Where B sits relative to the high of day and the premarket high. A B that is also the day's high adds a crowded level.
- Time of day. In FennEdge's Gapper HOD study, 70.2% of ≥20% gappers printed their 04:00–20:00 high before 10:00 ET (n = 671; 2026-07-01 to 2026-10-02; FennEdge archive; as of 2026-10-02; see Small-cap gapper statistics (2026)). A D projected above the morning high is a projection into a level that many such days did not revisit. That is context, not an ABCD statistic.
- Labelling agreement. Patterns two reviewers labelled the same way vs. patterns they disagreed on.
Cost of being wrong
Using the Pellucid Grid plan at 454 shares (Illustrative):
| Outcome | Exit | R after costs | Dollars after costs |
|---|---|---|---|
| Path A: target D | $3.12 | +3.68R | +$183.87 |
| Path B: stop under C | $2.58 | −1.23R | −$61.29 |
| Path C: time stop | $2.74 | +0.23R | +$11.35 |
| Path B, faster break with 6¢ slippage | $2.54 | −1.59R | −$79.45 |
Fees are about $2.27 per round trip at this size. With a stop under C, the risk per share depends on how deep C is; a deeper C means a smaller size for the same dollar risk.
How many trades before this means anything
Low-hit-rate, high-payoff plans need larger samples. With a true Path A share of 15%, the chance of seeing no Path A at all in 10 trades is 0.85¹⁰ ≈ 20%, and in 20 trades about 4% (computed from the assumed 15%, Illustrative). A trader who stops after 10 straight misses may be dropping a plan that is near break-even, or keeping one that is not; 10 trades cannot tell. With 20 labelled Cs and 3 reaching D (Illustrative counts), a 95% Wilson interval for the Path A share runs from about 5% to 36%. How many trades before your numbers mean anything explains why.
What this lesson does not claim
- It does not claim ABCD patterns reach D, or that trading them is profitable or unprofitable. The odds box is Illustrative, and no FennEdge study yet covers the pattern.
- The labelling rules are one workable choice; Fibonacci-based rules are equally valid choices with different counts.
- The measured move is a description of some completed sequences, not a target that price owes.
- Pellucid Grid, its bars, quote and costs are fictional and Illustrative.
Go deeper
- FennEdge first.
- In-Play Days shows 1-minute charts of past active days. Without Max, including as a guest, you can open each ticker's most recent active day in full; earlier days and the two-year history need Max (tiers as of 2026-10-05). Hide the chart after a chosen time, label A, B and C, then reveal; replaying many past days needs Max.
- Gapper HOD Timeline shows when each gapper's high printed (guests, Free and Plus: latest completed session; earlier sessions need Pro or Max), useful context for where B and D sit.
- Fibonacci retracements and extensions. The ratios (38.2%, 61.8%, 1.272, 1.618) come from older pattern literature and later harmonic-pattern writers. The pattern is often traced to Gartley, H. M. "Profits in the Stock Market" (1935). They are hard to test because so many levels exist that one usually fits.
- Labelling bias between reviewers. Measure agreement before measuring outcomes.
- Related lessons: Bull flags, The first pullback, Thinking in probabilities, Backtesting intraday ideas.
Exercises
Six exercises follow this lesson: three concept checks, two replays of the static sequence above (classify labels as rule-consistent, reviewer-dependent or inference; then pick every outcome consistent with the information at 09:59 before Path B is revealed), and one practice task. Exercises, answers and progress are available after signing in.
Practice task: label 20 sequences twice. On replayed charts (for example in In-Play Days), find 20 candidate ABCDs using your written rules, labelling each at the decision time with later bars hidden. A week later, or with a second person, label the same 20 again without looking at the first labels. Record both sets of labels, whether they agree, and the outcome of the plan from the agreed labels (Path A, B or C, R after costs with a slippage penalty). Log candidates where C broke before confirmation too. Use the FennEdge practice log: download the practice log (CSV), with a column guide. Keep it separate from the Trading Journal, which holds only real broker-synced trades.
Primary sources
- FINRA. Frequent Intraday Trading: Understanding the Basics.
- FINRA. Order types.
Sources checked 2026-10-05. Pellucid Grid and all prices, volumes, outcome splits and costs in this lesson are fictional and Illustrative. Measured figures come from the FennEdge archive as stated. This lesson is education, not a recommendation to buy or sell any security. Past patterns do not indicate future results.