Elliott Wave tells you the shape of a move; Fibonacci tells you its size. The two have been intertwined since Ralph Nelson Elliott himself noticed that the relationships between waves kept landing on the same handful of ratios — the very ratios derived from the Fibonacci sequence. That observation is what transforms wave analysis from an exercise in drawing squiggles into something measurable: with Fibonacci, you can project where a wave is likely to turn, estimate how far a trend might run, and define your invalidation in terms of concrete prices rather than vague feel. This guide lays out the ratios that genuinely matter and how to apply them to a live count.
It builds directly on Elliott Wave theory explained and pairs closely with Elliott Wave extensions, where the extension ratios are explored in more depth.
Key takeaways
Q: Why does Elliott Wave use Fibonacci ratios?
A: Ralph Nelson Elliott observed that the relationships between waves frequently align with the Fibonacci sequence and its ratios — especially 0.382, 0.618 and 1.618. The ratios give a way to measure and project where waves are likely to turn, rather than relying on shape alone.
Q: What Fibonacci retracement does wave 2 use?
A: Wave 2 commonly retraces 50% to 61.8% of wave 1, and sometimes as deep as 78.6%. Wave 4 is usually shallower, holding near the 38.2% retracement of wave 3, reflecting the guideline of alternation.
Q: How far does wave 3 extend?
A: An extended wave 3 commonly reaches 1.618 times the length of wave 1, and in very strong trends 2.618. When wave 3 extends, wave 5 often works out roughly equal to wave 1.
Where the ratios come from
The Fibonacci sequence — 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, and so on, each number the sum of the two before it — produces a set of constant ratios as it progresses. Divide any number by the one that follows and you approach 0.618; divide by the number two places ahead and you approach 0.382; the inverse of 0.618 gives 1.618, the famous golden ratio. These proportions appear throughout nature, and Elliott argued they appear in markets too, because markets are the aggregate expression of human behaviour and human behaviour is not exempt from the same patterns.
For a wave trader, the philosophical question of why markets respect these ratios matters less than the practical observation that, often enough to be useful, they do. The key levels worth committing to memory are the retracements — 0.236, 0.382, 0.5, 0.618 and 0.786 — and the extensions — 1.272, 1.618 and 2.618. Note that 0.5 is not strictly a Fibonacci ratio, but it is included by convention because the halfway retracement is so commonly respected.
Retracements: sizing waves 2 and 4
Retracements measure how far a corrective wave pulls back against the preceding move, and they are most useful for the two countertrend legs of an impulse: waves 2 and 4.
Wave 2 tends to be deep. It commonly retraces between 50% and 61.8% of wave 1, and in sharper corrections can reach 78.6%. The depth reflects lingering disbelief — after the first tentative move off a low, the crowd is not yet convinced, so the pullback is substantial. What wave 2 must never do is retrace more than 100% of wave 1; that is rule 1, and it gives every wave-2 entry a clean invalidation.
Wave 4 tends to be shallow. It usually holds near the 38.2% retracement of wave 3, occasionally reaching 50%. By the time wave 4 forms, the trend is widely accepted, so the market merely pauses rather than deeply retraces. This contrast between a deep wave 2 and a shallow wave 4 is the guideline of alternation expressed in numbers, and it is one of the most dependable tendencies in the whole framework.
Extensions: projecting waves 3 and 5
Extensions measure how far a wave travels beyond the length of a prior wave, and they are how you project targets for the trending legs.
Wave 3 is the leg that most often extends, and it commonly reaches 1.618 times the length of wave 1 — the golden-ratio relationship. In powerfully trending forex pairs it can stretch to 2.618. Measuring wave 1 and projecting these multiples from the end of wave 2 gives a reasoned target zone for where the strongest part of the move may finish.
Wave 5 has a few common relationships. When wave 3 has extended, wave 5 often works out roughly equal to wave 1 — the guideline of equality, and the single most practical wave-5 projection. Alternatively, wave 5 can relate to the net distance travelled from the start of wave 1 to the end of wave 3, frequently by 0.618 or 1.0. Having two or three candidate projections for wave 5 is normal; the value lies in seeing where they cluster.
The ratios are tendencies, not laws. Their power is not in any single level but in confluence — where several independent projections point to the same price. One Fibonacci level is a guess; three agreeing at one zone is a signal.
Confluence is everything
No experienced wave trader acts on a lone Fibonacci level. The market rarely respects a single number in isolation; it respects zones where multiple relationships agree. Suppose the 1.618 extension of wave 1, the 1.0 projection of wave 5 from a separate measurement, and a prior structural high all fall within a few pips of each other. That cluster carries far more weight than any one of those levels alone, because it represents several independent ways of arriving at the same conclusion.
Building confluence is therefore the real skill. Measure the retracements and extensions from more than one anchor, note where the projections stack up, and treat those clusters — not the individual lines — as the meaningful levels. When a confluence zone also lines up with a higher-degree wave boundary or a round number that the wider market watches, the case strengthens further. This is the same principle that governs the corrective patterns: a zigzag's C wave, for example, gains weight when its A=C projection lands inside an existing Fibonacci cluster.
Applying the ratios on forex
For currency traders, Fibonacci's appeal is that it needs no volume data — it is derived purely from price — so it works identically on decentralised spot forex as on any other market. The practical workflow is straightforward: once a count is structurally valid under the three rules, overlay the retracements on waves 2 and 4 and the extensions on waves 3 and 5, and check whether the labels and the ratios agree. When they do, confidence rises from "possible" to "probable"; when the ratios are wildly off, treat that as evidence the count may be mislabelled rather than evidence the market is misbehaving.
A word of discipline, though: Fibonacci levels are confirmation tools, not entry triggers in themselves. Price reaching a 0.618 retracement does not, on its own, mean a trade — it means a zone to watch for the structure to confirm a turn. Used that way, alongside the rules and the broader count laid out in how to count Elliott waves, Fibonacci becomes the measuring tape that makes the whole framework quantifiable rather than merely descriptive.
Common Fibonacci mistakes
Fibonacci is powerful precisely because it is simple, and that simplicity invites misuse. A handful of errors account for most of the trouble traders get into with it. The first is over-reliance: treating a Fibonacci level as a trade trigger in its own right rather than as a zone to watch for confirmation. Price touching the 0.618 retracement is not a signal; it is a place to look for the structure to confirm a turn. Acting on the level alone, without corroborating evidence from the wave count, is a fast route to being chopped up.
The second error is measuring from the wrong anchors. A retracement or extension is only meaningful if it is drawn from the correct swing high and low — the genuine start and end of the wave being measured. Sloppy anchoring, snapping the tool to an arbitrary candle wick rather than the structural turning point, produces levels that look authoritative but mean nothing. Because the whole value of Fibonacci is precision, imprecise anchoring quietly destroys it.
A third, subtler error is curve-fitting: cycling through every possible ratio until one happens to line up with where price already turned, then declaring Fibonacci "worked." Used this way, the tool can be made to fit anything in hindsight, which means it predicts nothing in advance. The discipline that prevents this is committing to the standard ratios and the standard anchors before price arrives, and treating a clean hit as confirmation rather than searching afterwards for whichever level happened to fit.
Finally, remember that 0.5 is not actually a Fibonacci ratio — it is included by convention because the halfway retracement is so often respected, but treating it as having the same theoretical weight as 0.382 or 0.618 misunderstands where the numbers come from. None of these mistakes is fatal on its own, but together they explain why some traders conclude Fibonacci "doesn't work" when the real problem is how they were using it.
Setting up the tools properly
In practice, two tools do most of the work: the Fibonacci retracement tool, for sizing the corrective waves 2 and 4, and the Fibonacci extension (or projection) tool, for targeting the trending waves 3 and 5. The retracement tool is drawn from the start to the end of the wave you expect to be retraced; the extension tool typically uses three points — the start and end of a reference wave plus the end of the correction — to project the next wave's likely length. Configuring both to display only the meaningful levels (0.382, 0.5, 0.618, 0.786 for retracements; 1.0, 1.272, 1.618, 2.618 for extensions) keeps charts readable and avoids the temptation to curve-fit against a dozen lines. Drawn carefully from genuine structural points, and read as confirmation rather than command, these two tools are all the measurement most wave analysis ever needs.
Wave 2 retraces deep (~0.618), wave 4 shallow (~0.382); wave 3 extends to ~1.618 of wave 1, and wave 5 often equals wave 1. The ratios are tendencies, and their real power is confluence. Anchor the tools to genuine structural points, commit to the standard levels before price arrives, and treat a clean hit as confirmation — never as a standalone trigger.
