Elliott Wave Theory
Elliott wave theory proposes that markets move in fractal cycles made of five impulse waves and three corrective waves. Murphy devotes chapter 13 of Technical Analysis of the Financial Markets to it. Elliott also held that successive waves tended to conform to Fibonacci ratios; tested on the Dow Jones from 1915 to 2003, the idea is not borne out (Batchelor and Ramyar, 2006).
Last updated: 20 October 2026
Prerequisites
PrerequisitesThis is one of the most advanced sheets in the track. It requires a command of Dow Theory (ch. 1, Easy level — the three nested trends), of trends and support/resistance (ch. 2, Easy level), and above all of Fibonacci retracements and extensions (ch. 5, Intermediate level), which are the principal measuring tool of Elliott waves.
Definition
DefinitionRalph Nelson Elliott (1871-1948), an accountant by profession, published his wave theory in The Wave Principle in 1938. From the late 1970s, Robert Prechter Jr. and A.J. Frost popularised it with Elliott Wave Principle (1978).
Murphy presents wave theory in chapter 13 (1999 edition, "Elliott Wave Theory"). Where Dow identifies three nested trends, Elliott proposes a fractal model: the same pattern repeats at every timescale.
Elliott's fundamental model consists of 8 waves:
5 impulse waves (in the direction of the main trend): waves 1, 3 and 5 are moving waves, waves 2 and 4 are corrective waves within the impulse. Within the theory, wave 3 is usually the longest and most powerful
3 corrective waves (against the trend), labelled A, B and C. The correction retraces part of the preceding impulse. Within the theory, wave C is usually at least as large as wave A
Why it matters
Elliott's theory offers an overarching framework: where most technical tools produce point signals — a crossover, a breakout, a pattern — it places every move within a cycle. For anyone using it, believing you are in wave 3 or in wave 5, at the end of a cycle, leads to opposite decisions.
Within the theory, the Elliott-Fibonacci connection is central. Prechter points out that the wave count lands on Fibonacci numbers (1, 1, 2, 3, 5, 8, 13, 21...) at every degree: one impulse = 5 waves, one correction = 3 waves, one cycle = 8 waves, then subdivisions of 34 waves, then 144, and so on. Those numbers follow from how the model is built (5 + 3 waves, each subdivided into 5 or 3): they do not show that prices respect Fibonacci ratios.
Elliott (1940) further held that successive moves, in price and in duration, tended to conform to Fibonacci ratios. Batchelor and Ramyar (2006) tested the idea on the Dow Jones from 1915 to 2003, comparing the ratios between successive moves with those of series drawn at random from the same index: the ratios occur no more often than chance, and the authors conclude that the idea can be dismissed. The example below shows a second difficulty: a count can be invalidated by the prices that follow, and must then be redone.
Key points
Elliott's three inviolable rules: (1) wave 2 NEVER retraces 100% of wave 1; (2) wave 3 is NEVER the shortest of waves 1, 3 and 5; (3) wave 4 NEVER enters the territory of wave 1. If any of these is violated, the count is wrong and must be revised
The Fibonacci targets the theory sets for each wave: wave 1 → no computable target, it is the starting point. Wave 2 → retraces 50% or 61.8% of wave 1. Wave 3 → extension of 161.8% or 261.8% of wave 1. Wave 4 → retraces 38.2% of wave 3 (alternating with wave 2). Wave 5 → often equal in length to wave 1, or 61.8% of the distance from wave 1 to wave 3
Fractality means the same 5-3 pattern repeats at every timescale. A wave 3 on the monthly chart contains a complete 5-3 cycle on the weekly, which itself contains a 5-3 cycle on the daily, and so on down
Concrete example
ExamplesThe S&P 500's bull market from 2009 to 2022 can be read as a complete Elliott cycle on the monthly chart. Wave 1 (March 2009 → May 2011, 667→1,370). Wave 2 (May 2011 → October 2011, a 21% correction in price, retracing roughly 42% of wave 1 in amplitude — between the 38.2% and 50% levels). Wave 3 (October 2011 → February 2020, 1,074→3,393, an extension of about 330% of wave 1). Wave 4 (February 2020 → March 2020, a 35% fall to 2,191, which stays above the top of wave 1: rule 3 holds). Wave 5 (March 2020 → January 2022, 2,191→4,818, the longest of the three impulse waves in points). The A-B-C correction that followed (January 2022 → October 2022, 4,818→3,491) retraced about 50% of wave 5 — Dow theory's 50% level. This count is reconstructed after the fact: a count ending wave 4 in December 2018 (2,346), at the low of the 2018 decline, is ruled out by the March 2020 low (2,191): wave 5 could not have started at 2,346, since price went lower before its January 2022 peak (rule 1 applied to its own sub-waves).
Common mistakes
CautionCount bias: the analyst forces the wave count to confirm a pre-existing opinion. Several counts can satisfy the three rules for a while (the example above sets aside one, ending in December 2018, that the March 2020 low invalidated): keeping them in view avoids settling on the one that suits
Treating Fibonacci ratios as a property of the waves. On the Dow Jones from 1915 to 2003, the ratios between successive moves fall on these values no more often than chance (Batchelor and Ramyar, 2006)
Ignoring the three inviolable rules. A count that violates any one of them is necessarily wrong, however elegant the scenario appears
Trying to count corrective waves in real time. Corrections are ambiguous by nature and can take many forms: a count is only confirmed once the correction is over, at the start of a new impulse
Sources
ReferencesBatchelor, R. and Ramyar, R. (2006), "Magic numbers in the Dow", Cass Business School, City University London. openaccess.city.ac.uk.
Elliott, R. N. (1938), The Wave Principle, and (1940), How the Wave Principle Works, and its Correlation with Mathematical Laws, reprinted in R. N. Elliott's Masterworks, New Classics Library, 1994, cited from Batchelor and Ramyar (2006).
Frost, A. J. and Prechter, R. R., Elliott Wave Principle, New Classics Library, first edition 1978.
Further reading
ProgressionFor comparison: Dow Theory (ch. 1) also describes nested moves, with its three trends and three phases, and Fibonacci ratios are covered in ch. 5. The three rules give each count a price level that invalidates it: one possible reference for placing a stop (risk management, ch. 9).
📊 Elliott Waves — The 5 + 3 Cycle
Market impact
MarketsRobert Prechter Jr. has published the newsletter The Elliott Wave Theorist since 1979, and some charting software offers wave-counting tools. None of the studies cited in this sheet measures an effect of Elliott counts on prices; the Batchelor and Ramyar (2006) test of the ratios between successive moves of the Dow Jones is negative.