Wyckoff Accumulation: Phases, the Spring, and What 33 Years of SPY Show

Michael Whitman · Trading since 2017, full-time since 2020 · YouTube creator, 137K subscribers · Published 2026-09-06 · Updated 2026-10-27 · Reviewed 2026-10-27

Covered on this page: TradingView.

Wyckoff accumulation is a schematic for the long sideways base that can follow a decline, divided into phases A to E and ending in a markup. Its best-known event, the spring, is a dip below the base that closes back inside; on SPY since 1993, 49 of 69 such dips closed below the low again within 10 sessions.

The phases can only be labeled with confidence once the advance has happened, which is why the schematic explains old charts far better than it identifies the present one.

How it works

A decline ends. Price moves sideways for a long time. Then it advances. The Wyckoff accumulation schematic describes that sideways stretch as a structured process rather than a pause.

The base is divided into phases A through E, each with its own named events — preliminary support, selling climax, automatic rally, secondary test, spring, sign of strength, last point of support. It is the most detailed vocabulary in technical analysis.

The organizing idea is the composite operator — imagine all the informed buying as one participant accumulating a position without pushing price up. It is a useful way to think about why a base might look the way it does, and it is explicitly a fiction. Wyckoff said so; a lot of modern material presents it as an observation.

The spring, and the problem with all of it

The spring is a dip below the base that recovers quickly — read as the last shake-out before the advance. It is the most-traded event in the whole schematic.

The test is a second, shallower dip that confirms it. Two events, a sequence, a story.

And here is the difficulty: a spring and a breakdown are the same picture. Price dips below the base. If it recovers it was a spring; if it does not it was the base failing. The label is assigned by what happened next.

That is true of every phase in the schematic. Phase C is only Phase C because Phase D followed. A base that breaks down was never accumulation — it gets relabeled as distribution or as a continuation pattern, and the schematic’s record stays clean because the failures leave under a different name.

Nothing in the framework is numeric. How long must the base be? How deep may a spring go? How much volume counts as a climax? All unspecified, which means the schematic can be fitted to a very wide range of charts by anyone who wants to.

A worked example

SPY, June 2026. On 26 Jun 2026 SPY traded down to $716.58, below $722.59, the lowest low of the 20 sessions before (set on 9 Jun), and closed at $728.99, back above it. That is the spring shape. Over the next 10 sessions the lowest close was $741.00, and the close 10 sessions later was $749.17: this one held.

SPY, September 2026. On 1 Sep 2026 SPY dipped to $759.48, below the 4 Aug low of $760.52, and closed at $761.78, back above. Same shape. Six sessions later, on 10 Sep, it closed at $757.83, below the level, and the close 10 sessions after the dip was $754.05. The first one gets called a spring in hindsight; the second gets called something else, and on the day both looked identical.

In practice: what is genuinely usable

The volume reading is the part worth keeping. Wyckoff’s core insight was to read price and participation together, and the observation that a base with declining volume behaves differently from one with rising volume is checkable and not derived from price. Checked here, it did not separate SPY’s springs, which is worth knowing before leaning on it.

Most modern Wyckoff material inverts the priority — heavy on phase labels, light on the volume analysis that the labels were originally supposed to summarize.

On a higher timeframe the entire schematic is a few flat bars. The phases need enough bars to be visible, so the framework lives at whatever resolution you chose.

A gap out of the base ends the structure without offering a price. The markup began overnight, and the last point of support never printed.

A failing base looks exactly like a working one until the moment it does not, which is the honest summary of the whole framework’s predictive content.

Each spring traded pays a full round trip, spread and commission, and a base that produces three dips before either resolving or failing charges for all three.

And the order book contains no phases. Nothing in the market’s data identifies who is buying or why. The composite operator is an interpretive device.

What Wyckoff accumulation is not

It is not observation of institutional buying. No feed identifies participants. The accumulation is inferred from price and volume behavior.

It is not a set of rules. It is a schematic with named events and no thresholds.

It is not distribution run backwards, exactly. Bases form slowly and tops form fast, so the two schematics have genuinely different rhythms even though they are drawn as mirror images.

And it is not falsifiable as commonly used. A base that fails is reclassified rather than recorded as a failure, which is the property to be most careful with.

How this connects

Accumulation is the general term; this is the structured version with named events inside it - a spring, a test, a sign of strength.

The Wyckoff upthrust is the mirror event at the top of a range, and the two are usually taught together.

And the phase is labeled afterward. A sideways range becomes accumulation once price leaves it upward, which is why the schematic looks cleaner than any live chart.

The original data

Sample: SPY daily bars from Yahoo Finance, 29 Jan 1993 to 26 Oct 2026. A dip is a session whose low trades below the lowest low of the 20 sessions before it, counted only if no dip happened in the 10 sessions before, and only from the 41st session (so 40 sessions of volume exist) with 10 later sessions available. A spring is a dip that closed at or above that 20-session low; a failure is a later close below it within 10 sessions. That gives 167 dips, from 2 Apr 1993 to 1 Sep 2026.

The finding: on SPY, a spring-shaped dip was more often the start of another close below the floor than a held low. 49 of the 69 springs (71.01%) closed below the level again within 10 sessions, at a median of 2 sessions and 13 of them the very next day; only 20 held. And SPY was higher 10 sessions later after 55.07% of springs, below the 60.72% of all sessions but within what chance produces on 69 cases (binomial p = 0.39), while the dips that closed below the low were followed by a higher close 66.33% of the time.

The volume check. Splitting the springs by whether average volume over the 20 sessions before was lower than over the 20 before that (the drying-up the schematic looks for): 41 springs came after volume dried up and 11 held (26.83%); 28 came without it and 9 held (32.14%). A Fisher exact test gives p = 0.79, so the difference is what chance produces, and the samples are small.

The median spring was shallow: 0.32% below the level. These dips are ordinary events in a market that spends most of its time below a prior high (the S&P 500 closed under an earlier record on 93% of trading days from 1950 to 2026), so a rule that fires on them fires often.

Second source. Nasdaq’s SPY bars from 26 Oct 2016 give 42 dips between 21 Mar 2017 and 1 Sep 2026, against 42 on Yahoo; 41 match on date and outcome (one falls on 10 Sep 2020 on Yahoo and 11 Sep 2020 on Nasdaq), both find 20 springs, and 20.00% held on both. Nasdaq’s stale 20 Apr 2026 row (open, high, low and close all equal, no volume) was dropped and its 9,999,999 volume on 17 Apr 2026 was treated as missing. Every dip is in the SPY spring file.

When it fails

In a range every dip below the floor that recovers is a spring by the definition. Ranges produce those regularly and go nowhere afterward, so the event fires most often exactly where it explains least.

The second failure is the base that never resolves. Months of sideways action with phases labeled along the way, capital committed, and no markup. Nothing in the schematic sets a time limit.

A third is the relabeling. When the base breaks down, the chart gets redescribed as distribution and the framework’s record is unaffected. Logging your own reads before the outcome is the only way to know what your hit rate actually is.

A fourth is skipping the volume. Phase labels without the participation analysis are the decoration without the substance, even though the simple drying-up test above did not separate outcomes.

And a fifth is applying the schematic to a fast chart. Wyckoff was describing multi-month structures in individual stocks. A five-minute accumulation schematic uses the vocabulary and none of the context.

Questions people ask

What is a spring in Wyckoff accumulation?

A dip below the base’s low that closes back inside it. On SPY since 1993, 69 dips below a 20-session low closed back above it the same day, and 49 of them closed below it again within 10 sessions.

What are the phases of Wyckoff accumulation?

Five, labeled A to E: the stopping of the decline, building the cause, the test (where the spring sits), the move to the top of the range, and the markup out of it.

Does volume confirm a Wyckoff spring?

Not in this test. Springs that followed falling volume held 26.83% of the time and the others 32.14%, a gap a Fisher test puts at p = 0.79.

How long does Wyckoff accumulation take?

The schematic sets no length. That is one reason it fits so many charts, and why each phase is easier to label after the markup than during it.

Related

Wyckoff method is the parent page with the full framework and its logic. Accumulation covers the concept without the schematic. And Wyckoff upthrust is the mirror event at the top of a range.

This page is educational, not financial advice. Test every idea on your own charts before risking money. Some links on this page earn a commission if you buy through them. It costs you nothing and it does not decide what appears here or in what order — how these pages are made is set out in our methodology.