For my first eight months of trading, I treated chart patterns like constellations: squint hard enough and you can see anything. I drew triangles on GBPUSD that existed only because I wanted them to, entered “breakouts” that were just Tuesday noise, and paid for every fantasy with real drawdown. The turn came on a grey Thursday in March, on a 100k evaluation I was quietly bleeding. On a messy GBPUSD M30 chart I saw a sharp sell-off, a bounce that stalled exactly where Fibonacci said it should, then a second leg down that stopped one pip inside the 127.2% extension of the bounce. I shorted the retest, rode it to the 61.8% retracement of the second leg, and banked 0.8R. Not luck — the first time I traded a structure with published, measurable rules instead of a vibe. That structure was the ABCD pattern, and it has stayed in my playbook ever since: on a funded account, you get paid for repeatable geometry with the risk maths attached.
In this article
- The GBPUSD Pattern That Changed How I Read Charts
- What the ABCD Pattern Actually Is
- The Complete ABCD Rule Set: Setup, Entry, Stop, Target
- The Fibonacci Maths Behind the Legs
- Worked Example: A Bearish ABCD on EURUSD
- I Tested It: 12 ABCD Trades on a 100k Demo
- When the ABCD Pattern Fails
- How the ABCD Pattern Fits Prop Firm Rules
- FAQ
The GBPUSD Pattern That Changed How I Read Charts
Here is the uncomfortable part: I almost deleted the pattern. Four swing points with letters on them looked like another fantasy — head and shoulders had failed me twice that month. But the numbers would not leave me alone: the BC bounce measured 61.8% of AB to within two pips, and the CD extension measured 127.2% of BC to within one. Randomness does not land inside Fibonacci tolerances twice on the same structure. That is the whole appeal of the ABCD pattern for a prop firm trader: it is the only common chart pattern whose validity you can check with a calculator before you risk a cent. Triangles are arguments. AB = CD symmetry is arithmetic.
Over the next four weeks I backtested the pattern by hand on EURUSD and GBPUSD — 60 occurrences — then traded it live on demo. Win rate: 58%, nothing spectacular. But the pattern told me exactly where I was wrong (above point C) and exactly where partials came out (38.2% and 61.8% of the CD leg). On a funded account, that precision beats a 70% win rate on vague setups. Below is the complete rule set: setup, entry, stop, targets, session filter, the Fibonacci maths with a worked example, a 12-trade test log, and the honest list of ways the pattern fails.
What the ABCD Pattern Actually Is
The ABCD pattern is a four-point harmonic structure: an impulse leg (AB), a corrective retracement (BC), and a second impulse leg (CD). In a bullish ABCD, A is a swing high, B a swing low, C a lower high, D a lower low — you buy at or near D expecting the correction of the CD leg. Bearish ABCD mirrors it: A is a swing low, B a swing high, C a higher low, D a higher high, and you sell at D.
What separates ABCD from the pattern zoo is that it is the root structure of the harmonic family — Gartley, bat, butterfly and crab are all ABCD cores with extra Fibonacci constraints bolted on. Learn one structure and you hold the engine of all of them. For prop firm purposes, one more property matters: the pattern completes at an extreme, so your stop has a natural home beyond the invalidation point and your first target is the retracement of the final leg. The geometry hands you the trade plan; your only job is the risk maths.
Compared with the patterns already on this site, ABCD differs from head and shoulders (a three-swing topping structure) and from trend pullbacks (which need a confirmed trend first). It appears inside trends, at trend ends, and in ranges — anywhere an impulse-correction-impulse sequence prints. The cost of that flexibility is discipline: enforce the ratio bands, or you are back to seeing constellations.
The Complete ABCD Rule Set: Setup, Entry, Stop, Target
My rule set, exactly as traded on funded accounts — every line exists because I lost money doing the opposite at least once.
| Rule | Bearish ABCD (sell at D) | Bullish ABCD (buy at D) |
|---|---|---|
| Setup legs | A = swing low, B = swing high, C = higher low, D = higher high | A = swing high, B = swing low, C = lower high, D = lower low |
| Timeframe | M30 or H1 only — M15 ABCDs complete inside spread noise on funded accounts | |
| BC leg | Must retrace 61.8%–78.6% of AB (accept 55%–85% with reduced size) | |
| CD leg | Must extend to 127.2%–161.8% of BC, or AB ≈ CD in length | |
| Entry | Sell stop 2 pips below D, or sell the first 15-min close back under D after a wick through it | Buy stop 2 pips above D, or buy the first 15-min close back over D after a wick through it |
| Stop loss | 1 pip above point C — the pattern is dead if C breaks | 1 pip below point C — the pattern is dead if C breaks |
| Targets | T1: 38.2% retracement of CD (take 50%). T2: 61.8% of CD (take 25%). Runner: trail below/above prior swing | |
| Session filter | Trade D completions only 07:00–10:30 GMT (London); skip 30 min either side of red-folder news | |
| Risk | 0.5% of account per trade, max 2 ABCD trades per day | |
Two rules deserve emphasis. First, stop above C (below C for bullish) is non-negotiable. Beginners place the stop beyond D, which is backwards: D is your entry zone, C is where the pattern’s logic dies. If price takes out C, you are holding a lottery ticket. Second, the news blackout matters more for ABCD than for most patterns — D-points love to print where news spikes land. A CPI candle can paint a perfect D and erase it. I skip the trade if high-impact news lands within 30 minutes of the D completion.
The Fibonacci Maths Behind the Legs
The ratio bands are the pattern’s quality control. Measure BC against AB with the Fibonacci retracement tool, then CD against BC with the extension tool. My acceptance table:
| Leg measured | Ideal ratio | Acceptable band | My action outside the band |
|---|---|---|---|
| BC vs AB | 61.8% retracement | 55%–85% | Below 55%: shallow pullback, skip. Above 85%: likely reversal, skip |
| CD vs BC | 127.2% extension | 127.2%–161.8% | Below 127.2%: incomplete, wait. Above 161.8%: overextended, skip |
| CD vs AB (length) | 100% (AB = CD) | 80%–130% | Way off: label it something else, do not force the letters |
| Time AB vs time CD | Roughly equal bars | CD between 0.5× and 2× AB’s bar count | CD much faster: exhaustion, halve size |
The formula side is simple enough to do on a phone calculator. Retracement percentage = (BC length ÷ AB length) × 100. Extension percentage = (CD length ÷ BC length) × 100. Position sizing for the trade uses the standard risk formula, and this is where the prop firm daily loss limit enters the maths:
Position-size formula (prop firm version):
Lots = Risk$ ÷ (StopPips × PipValuePerLot)
Risk$ = AccountSize × 0.005 (my 0.5% rule)
Sanity check: (Risk$ × MaxTradesPerDay) + open risk must stay well under your personal daily stop. I verify mine against the max daily drawdown calculator before every session.
On a 100k account: $500 risk, 65-pip D-to-C stop on EURUSD (1 lot = $10/pip) → lots = 500 ÷ (65 × 10) = 0.77. Two such trades risk $1,000 — 40% of my $2,500 personal daily stop. A 100-pip stop buys 0.50 lots with the same $500. The pattern sets the geometry; the formula sets the size. Never invert that order.
Worked Example: A Bearish ABCD on EURUSD
Tuesday 14 October, EURUSD M30, London session — walked exactly as my journal recorded it:
| Step | What happened | Numbers |
|---|---|---|
| 1. Legs form | A = 1.1720 swing low → B = 1.1640 swing high (overnight sell-off); bounce to C = 1.1689; second leg down to D = 1.1627 | AB = 80 pips; BC = 49 pips (61.3% of AB ✓); CD = 62 pips (127% of BC ✓) |
| 2. Ratio check | BC retracement inside 55–85% band; CD extension inside 127.2–161.8% band; CD bar count 9 vs AB bar count 11 — inside tolerance | All three checks pass — trade approved |
| 3. Entry + stop | Sell stop 2 pips under D; stop 1 pip above C. No red-folder news within 30 minutes — filter passed | Entry 1.1625; stop 1.1690; risk distance 65 pips |
| 4. Sizing | 100k account × 0.5% = $500 risk. 500 ÷ (65 × 10) | 0.77 lots → $7.70 per pip |
| 5. Targets | T1: 38.2% retracement of CD. T2: 61.8% of CD. Runner trailed under prior 15-min swing highs | T1 1.1601 (+24 pips); T2 1.1587 (+38 pips) |
| 6. Outcome | Price tapped 1.1629 (no fill on the wick — I used the stop order, not the touch), filled at 1.1625, hit T1 in 40 minutes, T2 into the New York open. Runner stopped at +31 pips | 50% at +0.37R, 25% at +0.58R, 25% at +0.48R → blended +0.45R = +$225 |
Notice the blended result: +0.45R, not the +2R Instagram version. ABCD’s first target sits close by design — the 127.2% extension completion means you enter at an extreme, so the first retracement target is shallow. The edge is frequency and precision of the stop, not home-run multiples. My 60-pattern backtest averaged +0.31R per occurrence with 58% reaching T1 — a grind-it-out edge, which is exactly what survives the daily drawdown environment: many small, well-defined risks, none capable of ending the account.
I Tested It: 12 ABCD Trades on a 100k Demo
After the backtest, I forward-tested the rule set on a 100k demo for four weeks (October 2026): 0.5% ($500) per trade, max two ABCD trades per day, London window only. The full log — losers included, because the losers are the point:
| # | Pair | Bias | Result (R) | Notes |
|---|---|---|---|---|
| 1 | EURUSD | Bearish | +0.6 | The worked example above |
| 2 | GBPUSD | −1.0 | C broke by 3 pips — stop did its job | |
| 3 | EURUSD | +0.4 | T1 only, reversed to entry | |
| 4 | USDJPY | +0.5 | Clean AB=CD, textbook | |
| 5 | GBPUSD | −1.0 | News spike through C — my blackout was 20 min, not 30. Fixed after this | |
| 6 | EURUSD | +0.9 | Runner caught the NY continuation | |
| 7 | EURUSD | +0.3 | T1 tagged, scratched the rest at entry | |
| 8 | GBPUSD | −1.0 | BC retraced 88% — outside my band, took it anyway. Lesson logged | |
| 9 | USDJPY | +0.7 | T1 + T2, runner stopped | |
| 10 | EURUSD | +0.4 | Slow grind, both partials | |
| 11 | GBPUSD | −1.0 | Valid setup, clean loss — this is the business | |
| 12 | EURUSD | +1.2 | Best of the run: D at London open, full extension to prior day low |
Totals: 8 wins, 4 losses (67%), +1.0R net = +$500 on the 100k demo at $500 per R. Worst day: −$1,000, inside my $2,500 personal daily stop. Average winner +0.63R, average loser −1.0R — the classic ABCD profile: full-R losses because the stop is structural, fractional-R wins because T1 sits close. I ran these through the risk of ruin calculator: at 67% win rate, 0.63:1 payoff and 0.5% risk, ruin probability on a 5%/10% drawdown account rounds to effectively zero across 500 trades. That — not the +$500 — is the real result: the maths cannot kill the account.
When the ABCD Pattern Fails
Five failure modes, each of which I have paid tuition for:
| Failure mode | What it looks like | The fix |
|---|---|---|
| C breaks before entry | Price rallies through point C while you wait for D — the structure is now a failed breakdown, not an ABCD | Cancel the pending order. A broken C voids the pattern, full stop |
| CD overshoots the band | CD extends past 161.8% of BC — momentum is stronger than the pattern assumed | Do not chase the extension. Let it go; an overextended CD usually means trend day, not reversal |
| News prints the D | A red-folder release spikes price exactly to the extension level, then reverses the spike | 30-minute news blackout either side. Trade #5 in my log is the receipt |
| Chop grinds through D | Low-volume Asian session drifts through the extension zone without conviction | Session filter: London window only. D-points printed before 07:00 GMT get watched, not traded |
| Forced letters | BC retraced 88%, CD hit 190% — you label it ABCD anyway because you want a trade | The ratio bands are the strategy. Outside them, there is no trade. Trade #8 is the receipt |
The meta-lesson: ABCD fails gracefully when you respect the invalidation, catastrophically when you negotiate with it. The stop above C is small by design — 65 pips on 0.77 lots in my example, a $500 risk on 100k. Four straight clean losses cost $2,000, still inside the daily stop. The pattern only turns dangerous when a trader widens the stop “to give it room,” converting a measured 1R risk into an undefined one. On a funded account, undefined risk is the breach.
How the ABCD Pattern Fits Prop Firm Rules
This is the section that decides whether a strategy belongs on this site. A pattern can be beautiful and still unfundable — if it needs 3% risk, holds through news, or fires 12 times a day, it breaches a challenge account regardless of the geometry. ABCD passes the prop firm audit on four counts.
First, risk is structurally capped. 0.5% risk per trade with a stop above C means the worst outcome is known before entry — and the consistency rule (no single day above 30–40% of total profit) is easy to satisfy, because no ABCD trade produces a monster day. My best test trade made +1.2R; nothing here flags a consistency review.
Second, frequency is naturally throttled. Clean ABCDs print 2–4 times a week per pair on M30/H1. You cannot overtrade a pattern this selective — and since overtrading ends more challenges than bad analysis, that is a feature.
Third, it is fully manual and rule-based — clear of the EA and trade-copier restrictions many firms attach to evaluations. A Fib tool, a pending order, a journal. Nothing to code or VPS-host.
Fourth, the session filter and D-point confirmation entry (the 15-minute close back through D, not a blind limit at the touch) keep you out of low-liquidity traps. And the news blackout doubles as compliance if your firm restricts news trading: you never hold a fresh ABCD position into a red-folder release.
FAQ
What is the ABCD pattern in trading?
It is a four-point harmonic structure: an impulse leg (AB), a retracement (BC, ideally 61.8% of AB), and a second impulse leg (CD, ideally a 127.2%–161.8% extension of BC). Traders enter at or near point D, expecting price to retrace the CD leg. It is the core structure behind Gartley, bat and butterfly patterns — The Trading Bible’s ABCD guide has a good visual walkthrough of the classic rules.
What are the exact Fibonacci ratios for a valid ABCD?
BC should retrace 61.8%–78.6% of AB (I accept 55%–85% with reduced size), and CD should extend 127.2%–161.8% of BC. The classic textbook version also requires AB ≈ CD in length. If any leg falls far outside its band, it is not an ABCD — do not force the letters.
Does the ABCD pattern work on the M15 timeframe?
On a funded account, I do not trade it below M30. On M15 the extension levels get tagged by spread widening and session noise, and the stop ends up so tight that normal volatility takes you out first. M30 and H1 are the funded-account timeframes for this pattern.
Should I enter exactly at point D or wait for confirmation?
I use a stop order 2 pips beyond D, or I wait for a 15-minute close back through D after a wick. Blind limit orders at the extension level get filled on wicks that never confirm — trade #1 in my log filled on the stop order, not the touch, and that distinction saved me from two fake-D losses in the backtest.
How is ABCD different from head and shoulders?
Head and shoulders is a three-swing reversal with a neckline; ABCD is a measured-move harmonic with Fibonacci-defined legs and no neckline. ABCD also prints far more often — inside trends and ranges — while head and shoulders needs a mature trend to reverse.
Can I automate the ABCD pattern on a prop firm account?
Technically yes, practically be careful: many firms restrict EAs during evaluations, and ABCD needs discretionary ratio checks that are awkward to code. I trade it manually — Fib tool, pending order, journal. Manual keeps you clear of automation clauses and forces the ratio discipline that makes the pattern work.
Risk disclaimer: Trading forex, CFDs and other leveraged products on prop firm or personal accounts carries a high risk of loss. Nothing here is financial advice — the examples use a demo account and past pattern behaviour does not guarantee future results. Always verify your firm’s current rulebook before trading.





2 comments
[…] is this different from the ABCD pattern strategy I covered earlier? The ABCD is a four-point, three-leg pattern with no X leg and no […]
[…] ABCD pattern […]