Price Action vs Elliott Wave
Michael Whitman
·
Trading since 2017, full-time since 2020
·
Published 2026-09-09
·
Updated
·
Reviewed
Price action reads the bars and swings themselves with no model of how markets must move. Elliott wave applies a fixed model, five waves with the trend and three against, at every scale, and checks each count against three rules that a valid impulse may not break.
Price action reads the bars with no theory of what must come next. Elliott wave brings a complete theory and fits each chart into it. The useful difference is that Elliott’s theory has hard rules, and rules can be counted. Run over seventy years of the S&P 500, they rejected most five-swing stretches at any single scale, and still left a valid-looking count around most sessions once the scale was allowed to change.
What each one is
Price action reads the chart directly - where bars closed, which highs and lows held, how swings compare - with no overarching model. Price action covers the practice.
Elliott wave applies a model: markets advance in five waves and correct in three, and the same pattern repeats at larger and smaller degrees. Elliott wave covers the counts. StockCharts credits the theory to R.N. Elliott in the 1930s and its popularity to Robert Prechter in the 1970s.
Elliott’s model comes with three rules for an impulse. Wave 2 cannot retrace more than all of wave
- Wave 3 can never be the shortest of waves 1, 3 and 5. Wave 4 can never overlap wave 1’s price territory. A count that breaks one is not an impulse and has to be relabeled.
Where they differ
Whether a reading can be checked. A price-action read such as “buyers defended the low” has no rule attached, so it cannot fail a test. An Elliott count can: the three rules either hold for the five swings or they do not.
What happens after a miss. A wrong price-action read is simply wrong. A failed Elliott count can be moved to another degree, where a different set of swings is tested, so the framework survives the miss.
How much the reader must choose. Price action asks you to choose which bars matter. Elliott asks you to choose the swings, the starting point and the degree, three choices that each change the answer.
What each one needs to be useful. Price action needs a written level supplied by the reader. Elliott supplies its own levels, the start of wave 1 and the top of wave 1, and needs the reader to honor them.
Counting the rules by machine
Swings first. A zigzag on daily closes keeps a turn only once price has moved a set percentage away from it; 5% is the traditional setting, and 3% and 10% were run as smaller and larger degrees. Every run of five consecutive zigzag legs starting at a low was checked as an upward impulse, and every run starting at a high as a downward one.
Then the three rules, exactly as written. Wave 2 ends above wave 1’s start. Wave 3, measured in index points, is not shorter than both wave 1 and wave 5. Wave 4 ends above wave 1’s top. The mirror applies to downward counts.
What this can and cannot say. It counts how often swing sequences satisfy the rules. It does not apply Elliott’s many guidelines, diagonals or corrective patterns, and it does not ask what came next.
A count that passed, then met a new high
The 5% zigzag’s most recent passing count starts at the S&P 500’s 8 Apr 2025 close of 4,982.77. Wave 1 ended a day later at 5,456.90. Wave 2 ended on 21 Apr at 5,158.20, above the start. Wave 3 ran to 6,890.89 on 28 Oct 2025, the longest of the three. Wave 4 ended on 20 Nov at 6,538.76, above wave 1’s top. Wave 5 topped at 6,978.60 on 27 Jan 2026. All three rules hold.
Then the chart kept going. The next zigzag low came on 30 Mar 2026 at 6,343.72, and before 2 Oct 2026 the index closed as high as 7,798.99, above the end of the supposedly complete five waves. Under the model that is not a failure; it means the count belongs inside a larger degree, which is the flexibility the next section measures.
Which one to use: Price Action or Elliott Wave?
Use price action when you want a reading tied to what the bars did, with a level you choose and a level that would prove you wrong. It claims less and is easier to keep honest.
Use Elliott wave when you will commit to the degree and the invalidation before the move. Written down that way, the three rules make the count a genuine claim; chosen afterwards, almost any chart offers one.
Use price action when you cannot hold that discipline, because a loose read of a loose method is less misleading than a loose read of a rigid one.
And use both rules of thumb from each: the price-action habit of reading closes against levels, and Elliott’s habit of naming in advance the price that ends the idea.
Why changing the degree rescues a count
At any single swing size, the rules rejected about nine sequences in ten. At 5% swings, 42 of 404 five-leg runs passed; at 3%, 73 of 881; at 10%, 11 of 102.
Across the three sizes together, the passing counts covered two-thirds of the calendar. 11,664 of 17,514 sessions (66.60%) sat inside at least one rule-passing upward count at 3%, 5% or 10%. A reader free to switch degree will usually find a valid five-wave advance around any date, which is why a count stated without its degree tests very little.
The original data
Sample: S&P 500 daily closes from 4 Mar 1957, the day the 500-stock index began, to 2 Oct 2026, 17,514 sessions (Yahoo Finance, ^GSPC). The index closed higher on 9,291 of 17,513 day-to-day changes (53.05%), lower on 8,148 (46.53%) and unchanged on 74.
| Zigzag size | Turns | Median leg | Five-leg runs | Passed all three rules | Wave 4 rule held |
|---|---|---|---|---|---|
| 3% | 886 | 11 sessions, 6.00% | 881 | 73 (8.29%) | 121 |
| 5% | 409 | 23 sessions, 9.39% | 404 | 42 (10.40%) | 53 |
| 10% | 107 | 69.5 sessions, 19.33% | 102 | 11 (10.78%) | 18 |
Upward counts passed more often than downward ones at every size: 50 of 441 against 23 of 440 at 3%, 29 of 202 against 13 of 202 at 5%, and 6 of 51 against 5 of 51 at 10%, which fits an index that rose over the period. The wave-4 rule was the hardest to meet: at 5% swings only 53 of 404 runs kept wave 4 above wave 1’s top (or below its bottom on the way down).
Coverage by degree: sessions inside a rule-passing upward count were 37.75% at 3%, 49.21% at 5% and 35.90% at 10%, and 66.60% at one or more of the three.
What this does not show. Nothing here measures whether a passing count is followed by anything in particular; the counts are found in hindsight, from confirmed swings. Every five-leg run, its six turning values and which rules it passed are in the S&P 500 zigzag rule-check file.
When it fails
The characteristic price-action failure is a read that cannot be wrong. “Buyers are stepping in” and “structure is holding” fit whatever comes next because they name no price. A trader can do this for years, feel that the chart makes sense every day, and have no record that could show otherwise. The fix is the same in both methods: a written level, and a written price at which the read is abandoned.
The characteristic Elliott failure is the recount. When the market goes past the end of a complete count, as it did after 27 Jan 2026, a new degree makes the count valid again and the test disappears.
A third failure, shared by both, is changing timeframe until the chart agrees.
A fourth is reading the swings before they are confirmed; a zigzag turn is only known after the move away from it.
And a fifth is treating either as a full method; neither sets a position size.
Related
Price action covers reading bars and swings without a model. Elliott wave covers the waves, the degrees and the rules counted here. And technical analysis covers the wider chart-reading tradition both methods belong to.
Whichever you use, write the reading down before the move with the price that would prove it wrong. For an Elliott count that also means writing the degree, because a count that may move to another degree after the fact can absorb almost any chart.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.