Price Action vs Elliott Wave: Where Each Says You Are Wrong, on Real SPY Data

Michael Whitman · Trading since 2017, full-time since 2020 · YouTube creator, 137K subscribers · Published 2026-09-09 · Updated 2026-11-02 · Reviewed 2026-11-01

Price action and Elliott wave look at the same highs and lows. Price action stops at describing them; Elliott wave fits them into a five-up, three-down model and forecasts the next leg. The practical difference is where each one tells you that you are wrong. Price action draws that line at the last higher low. A valid Elliott count draws it lower, at the top of wave 1, and offers more than one count to choose from.

This page puts both methods on the same 33 years of SPY swings, reads the October 2026 chart both ways with real prices, and measures how far apart their lines sit and whether either signal said anything about the next 60 sessions.

What each one is

Price action is reading the chart as it is: which swing highs and lows held, whether the newest ones are higher or lower than the last, where closes landed. It makes no promise about the next leg. The full practice is on the price action page, and the trend version of it is market structure.

Elliott wave starts from those same swings and labels them: waves 1 to 5 with the trend, A to C against it, nested inside larger and smaller waves. R.N. Elliott laid it out in the 1930s; Robert Prechter made it popular in the 1970s. Its three rules (wave 2 stays above the start of wave 1, wave 3 is never the shortest, wave 4 stays out of wave 1’s range) are covered on the Elliott wave page, which also tests each rule on its own.

So the honest framing is not two rival schools. Elliott wave is one specific way of reading price action, with a forecast added and a stricter test attached.

Where they differ

The core trade-off is claim versus room. Price action says less, so it is rarely without a reading and its line sits close to price. Elliott says more, which is why it gives a reading less often, and when it does, its line usually sits further away, while a second legal count can move that line again.

A worked example: SPY from April 2025 to October 2026

The swings come from a 5% filter on daily highs and lows, the setting behind a standard zigzag indicator: a swing is fixed only once SPY moves 5% the other way. From the 21 April 2025 low of $508.46, the filter found six more turns:

Since then SPY has made a high of $781.62 on 6 October 2026 and closed at $755.72 on 30 October, with a low that day of $750.36, 4.00% under the high. That is short of the 5% a new swing needs.

The price-action reading. By 9 March 2026 SPY had a higher high ($697.84 over $689.70) and a higher low ($650.85 over $508.46), so the read was an uptrend with its line at $650.85. The close of $648.57 on 20 March went through it, and the low of $629.28 that followed made a lower low. With highs rising and lows falling, the read was mixed from 8 April (the day that low was fixed) to 6 July, when the $716.58 low was confirmed above $629.28. Since then the reading has been up again, and the one price that ends it is a close below $716.58, 5.18% under the 30 October close.

The Elliott readings. The same swings support three different stories:

• Count 1, the swings as they come. Label $508.46 as the start, $689.70 as wave 1, $650.85 as wave 2, $697.84 as wave 3 and $629.28 as wave 4. Wave 4 sits $60.42 below wave 1’s top, so rule three is broken and this count is dead.
• Count 2, a larger degree. Treat the October-November 2025 dip as part of wave 1, so wave 1 runs from $508.46 to $697.84. Then $629.28 is wave 2, $760.40 is wave 3 and $716.58 is wave 4, $18.74 above wave 1’s top. All three rules hold, wave 5 is under way, and the count dies below $697.84, 7.66% under the last close. Wave 3 ($131.12) is shorter than wave 1 ($189.38), so rule two also caps wave 5: it cannot run past $847.70 without making wave 3 the shortest.
• Count 3, a fresh start. Begin at the $629.28 low instead: $760.40 is wave 1, $716.58 is wave 2, and wave 3 is now under way. This count survives anything above $629.28, 16.73% under the last close.

What the example shows. A price-action trader holding SPY has one level and no forecast. An Elliott trader has to pick a count first, and the pick decides both the story (last leg of an advance, or the middle of a new one) and the exit, which sits either $18.74 or $87.30 below the price-action line.

Price action is not free of choices either. On the five-bar swings used by the break of structure page, the 30 October close of $755.72 had already fallen below the 1 October swing low of $758.79, so a shorter-term reader would call the trend broken while the 5% reader still calls it up.

The original data

The test. SPY daily bars from 29 January 1993 to 30 October 2026, 8,497 sessions, raw prices, with the same 5% swing filter (402 swings) that the Elliott wave page uses. Both readings use only swings already confirmed on the day, so nothing looks ahead.

• Price-action reading: up when the last two swing highs and the last two swing lows both rose, down when both fell, mixed otherwise. Its line in an uptrend is the last swing low; a close below it while the reading is up is a trend break.
• Elliott reading: a live count exists from the day a wave 4 low is confirmed, if the five swings so far keep wave 2 above the start and wave 4 above wave 1’s top. It stays live until the 5% filter confirms the top of wave 5. Its line is wave 1’s top.

How often each had something to say. The price-action reading was up on 4,654 sessions (54.77%), down on 1,719 (20.23%) and mixed on 2,043 (24.04%); the first 81 had too few swings. A live Elliott fifth wave existed on 882 sessions (10.38%), from 18 counts, and 877 of those sessions were also price-action uptrends. Every live count passed rule two as well; in 2 of the 18, rule two capped wave 5 because wave 3 was shorter than wave 1.

How far away each line sat. On the day each count went live, the price-action line (the wave 4 low) was a median 4.64% below the close, in a range of 2.61% to 5.39%. The Elliott line (wave 1’s top) was a median 8.47% below, from 3.16% to 17.96%. It was the wider of the two in all 18 counts.

Which line got touched. While the 18 counts were live, a close went below the price-action line in 3 of them, each time on the very day the filter confirmed the fifth wave over. All three were “truncated” fives, where wave 5 topped below wave 3. The Elliott line was never reached while a count was live. Both results are mostly built into the test: both lines sat at least 4.76% under the high (wave 4’s low is confirmed only by a 5% rise), and a 5% drop from the high ends the count. Across all 18, wave 5 made a new high above wave 3 in 13 and fell short of it in 5.

What came next. Price action’s trend break and Elliott’s finished five are the closest thing each has to a sell signal. SPY closed higher 60 sessions after 10 of the 14 price-action breaks (71.43%) and after 14 of the 18 finished fives (77.78%), against 5,919 of 8,437 sessions of any kind (70.16%). Neither gap is distinguishable from chance (binomial p = 1.00 and p = 0.61), and both samples are small.

Degree is where Elliott gets its room. Looking back over the whole chart with three swing sizes, the stretches inside a rule-passing five covered 35.53% of sessions at 3%, 38.53% at 5% and 51.04% at 8%, and 69.70% (5,922 of 8,497) at one size or another. A reader free to change the swing size after the fact can find a valid count around most dates. The price-action reading has the same weakness at any one size, but it offers only one reading per size.

Every count, its five prices, both lines and what followed, plus every price-action break, is in the SPY price action vs Elliott file; the rule-by-rule check of all 198 five-swing runs is in the SPY Elliott rule file.

Which one to use

Use price action when you want a reading on most days and an exit close to price. It answers one question, whether the swings are still rising, and the answer changes only when a close crosses a level anyone can see.

Use Elliott wave when you will commit to a count, its degree and its invalidation price before the move. Written that way it adds two things price action lacks: a stage (is this the third leg or the fifth?) and rule two’s ceiling on wave 5. Chosen after the fact, as the degree test shows, it can fit almost anything.

Do not use either as a timing signal on its own. In this sample, neither SPY’s trend breaks nor its completed fives changed the odds of SPY being higher 60 sessions later.

Using both together

Common forum advice is to let Elliott wave set the direction and price action set the entry. The data above suggests a narrower job for each. Elliott’s useful output is the list of levels where a count dies, written beside its alternative; price action’s useful output is the nearest level where the trend reading flips. In the October 2026 example that means three numbers on one chart: $716.58 for the trend, $697.84 for count 2 and $629.28 for count 3. Price crossing the first one tells the price-action reader to step aside, while the Elliott reader still has to see which count survives.

When it fails

Price action: the read that bends to fit

“Buyers are defending this area” names no price, so it cannot be wrong. Price action only becomes testable when the level is written down, as the $716.58 line is above.

Price action: swing size changes the answer

On 30 October 2026 the five-bar reading had flipped down while the 5% reading was still up. Neither is wrong, so choose the size before you look, not after.

Elliott: the recount

A count that breaks a rule can be relabeled at another degree, the way count 1 above becomes count 2. That is allowed by the theory and it is also why a count stated without its degree claims very little.

Elliott: the wide exit

The line that kills a fifth-wave count sat a median 8.47% below price in this sample, nearly double the price-action line. Sizing off the Elliott line means accepting a much larger loss before the idea is proven wrong.

Questions people ask

Is Elliott wave a type of price action analysis?

Yes. It uses nothing but price swings, so it sits inside price action, with a model of how those swings should unfold added on top.

Can you combine price action and Elliott wave?

You can, and many traders do. Treat the Elliott count as a set of written scenarios with their own invalidation prices, and let a close through the nearest price-action level decide when to act.

Which is better for beginners?

Price action. It has one reading per swing size and one line to watch; Elliott wave needs that skill first, then labeling, degrees and the three rules on top. The step-by-step labeling process is in how to count an Elliott wave.

Does Elliott wave predict better than price action?

Neither predicted anything in this test. SPY’s 60-session results after a price-action trend break and after a finished Elliott five were both in line with any random day.

Related

Price action covers reading bars and swings without a model, and market structure turns those swings into the up, down or mixed reading used here. Elliott wave explains the waves, the degrees and the three rules, with the rule-by-rule test on the same SPY swings. Swing highs and lows decide which turns count at all. Technical analysis is the wider tradition both belong to.

This page is educational, not financial advice. Test every idea on your own charts before risking money.