What it is
The ABCD pattern is the simplest continuation structure on any chart: an impulse leg (A to B), a pullback (B to C) that retraces part of the impulse, and a second leg (C to D) that is expected to be roughly the same size as the first. "Measured move" is the same idea applied to any pattern: the projected target equals the size of the move that preceded the consolidation. This playbook enters near C and targets D.
It is a framework for the question "how far might this go", which is otherwise where most discretionary traders guess.
The logic
Markets move in impulses and corrections because the participants who drive a move do not execute all at once. A fund building a position buys, pauses to avoid pushing price, then buys again; the second leg is often similar in size because the same participant, with the same urgency, is doing the same thing. The pullback also shakes out weak hands whose exits are absorbed by the patient buyer.
On the other side are traders who sold B as a top, and traders who shorted C expecting a full reversal. When the second leg starts, those shorts cover and add to the move. When the pattern fails, price breaks C and the pattern becomes the first leg of a reversal for the other side.
Setup rules
- Market: any liquid market; the pattern is scale-independent. Intraday it works best on 5 and 15-minute charts of index futures and large caps; on daily charts it becomes a swing-trading setup.
- Impulse (AB) conditions: at least 1.5 ATR of the chosen timeframe, made of mostly one-directional candles, with rising relative-volume.
- Pullback (BC) conditions: retraces 38 to 62 percent of AB (see fibonacci-retracement); volume contracts; no close below the 50 percent level on the timeframe you entered on; takes fewer bars than AB.
- Context: trade only in the direction of the higher-timeframe trend. An ABCD long against a daily downtrend is a lower-probability trade.
Entry, stop, target
Enter when price breaks the high of the pullback's last bar (a change of character), or more aggressively at the 50 to 62 percent retracement with a limit order. The stop goes just below C. The target D is C plus the length of AB; a conservative version targets C plus 0.618 of AB.
| Item | Level | Notes |
|---|---|---|
| A | 100.00 | Start of impulse |
| B | 104.00 | AB = 4.00 |
| C | 102.00 | 50 percent retracement |
| Entry | 102.60 | Break of pullback bar high |
| Stop | 101.70 | Below C, risk 0.90 |
| Target D | 106.00 | C + AB, reward 3.40, about 3.8R |
| Conservative D | 104.47 | C + 0.618 AB, about 2.1R |
Scale out a third at the conservative target and the rest at D, or trail behind the last swing low on the entry timeframe once price passes B.
Position sizing and risk
Because the stop sits at a structural level (below C), it is often further than a fixed-tick stop, and the size must shrink accordingly. Compute it from the stop distance at /tools/position-size, and hold risk per trade at 0.5 to 1 percent following /learn/risk-management. Do not move the stop tighter to justify a larger size; that converts a structural trade into a coin flip.
What breaks it
- Symmetry is a tendency, not a law. Second legs are frequently shorter (a failed measured move) or much longer (a trend day). The 3.8R in the table is an upper bound, and realized expectancy depends heavily on scaling rules.
- Deep retracements. A pullback beyond 78 percent is usually the beginning of a reversal, and entries there are guessing.
- Chop. In a range every impulse is followed by a full retracement, so every ABCD fails at C. Trend filters are essential.
- Costs and slippage are small relative to the structural stop, which is one reason the pattern remains useful.
- Subjectivity. Two traders will mark different A and B points. Fix your rules for what counts as an impulse or the backtest is meaningless.
- Timeframe mismatch. A clean ABCD on the 5-minute chart that projects D into a daily-chart resistance level will usually stop short of it. Check the higher timeframe before trusting the projection, and take the conservative target when D sits beyond a visible level.
How to test it
Define the impulse and pullback algorithmically (for example, an AB leg is a swing of at least 1.5 ATR with no counter-close larger than 0.3 ATR) so a script can find it. Run it across 5 years of daily data on a broad universe or 12 months of 5-minute data on a few futures. Record where D landed relative to the projection as a ratio; the distribution of that ratio is the real edge (or absence of it). Minimum 300 patterns. Then trade 50 in a simulator to confirm you can recognise them live.
Variations
- Three-drive pattern: a third leg after another pullback; targets tend to degrade with each leg.
- Flag measured move: the ABCD applied to a flag-pattern where BC is a tight sideways channel.
- Failed ABCD reversal: short a break of C with a target back at A.
Further reading
fibonacci-retracement, pullback, trend, higher-highs-higher-lows, flag-pattern, atr, r-multiple, confluence, timeframe.
Related playbooks: ema-pullback-trend, three-day-pullback, opening-range-breakout, base-breakout