Wyckoff phases — fixed ranges, alternatives and label agreement

Wyckoff’s framework can organise observations of a range, its boundaries and later movement outside it. Calling a region accumulation or distribution remains an interpretation. A chart alone does not reveal bank portfolios, participant identities or the purpose of their orders.

Pair each interpretation with a predeclared rejection condition. “I am considering accumulation, but a Close below the fixed lower boundary rejects this version” is testable. “Institutions are collecting positions here” is not established by candle shape alone.

Distinguish the cycle from the A–E range phases

The idealised cycle uses accumulation, markup, distribution and markdown. The trading-range schematics separately use phases A–E. Wyckoff Analytics calls the Composite Man a heuristic device, rather than an identified single FX institution. Continuation ranges can be interpreted as reaccumulation or redistribution.

Preserve the original range, drawing time and later revisions. Different boundaries can yield different phase labels. A range with no resolved interpretation is a valid observation; a label that needs later candles is not information available earlier.

Accumulation: measure before assigning meaning

The terminology includes preliminary support (PS), selling climax (SC), automatic rally (AR), secondary test (ST), sign of strength (SOS) and last point of support (LPS). These are schematic descriptions, not a fixed series that every pause must complete within a particular month-count.

Keep separate fields for prior direction, boundaries, closing prices and an alternative explanation. A lower wick has length min(Open, Close) − Low. Measuring it does not establish who bought or whether those buyers retain positions. Do not move support lower after each new decline merely to preserve a preferred label.

Distribution: preserve the competing continuation case

Distribution terminology includes preliminary supply (PSY), buying climax (BC), automatic reaction (AR), secondary test (ST), sign of weakness (SOW) and last point of supply (LPSY). A post-rise consolidation may also precede another rise or repeated range crossings. An upper wick does not independently settle the outcome.

Write down which future observation would weaken your interpretation before seeing it. A rule that calls the same return favourable for both incompatible directions does not distinguish them. The schematic is not itself a complete short-entry or sizing specification.

Spring and UTAD are optional events

A spring breaches the lower range boundary then returns; UTAD breaches the upper boundary in the distribution schematic. The teaching source includes versions without these events. A no-spring/no-entry rule can describe a chosen strategy, but does not define every Phase C.

Specify whether return requires a quote, a touch or a completed Close and set any allowed delay before testing. There is no universal pip allowance or one-to-four-bar rule inherent in the name. The eventual Close was not known at the earlier intrabar breach.

The diagram’s O = 105, H = 106, L = 101, C = 104 meets a breach-and-Close-return rule for a previously frozen 102–110 range, but not 100–110. Do not relocate the original lower boundary after seeing Low 101. Begin with reproducible support and resistance construction and the broader technical-analysis workflow.

Two separate fictional scenarios with one width

Use two independent Bid EUR/USD scenarios, both with 1.1000–1.1100 boundaries fixed before t1. Earlier anchor data is assumed, not reconstructed here. Pip is 0.0001, so the range width is 100 pips. This is a demonstration, not a historical sample or proof of an identified phase.

At t1, each breach is 20 pips, or 20% of the width. The two Closes return 30 pips above the lower or below the upper boundary. At t2, Close is 20 pips beyond the boundary in the considered direction. At t3, A’s Low is 5 pips above former resistance; B’s High is 5 below former support. Those later observations were unavailable at t1.

The last row is an alternative future, not an obligatory fourth step. For this model choose a Close strictly beyond the opposite boundary to reject the version; equality fails that condition. These are demonstration rules, not universal recommended tolerances. No Ask, fills, commissions, financing or exit rule is supplied, so the distances are not trade profits.

Define the volume field and check label agreement

MetaTrader describes Forex tick activity as a price-change count. BIS describes fragmented OTC execution. One feed does not cover every transaction; a higher counter does not prove institutional absorption or breakout direction. Record feed, interval, clock and units; see Forex volume measurement.

Before testing trading results, check recognition consistency. Two independent raters label ten fictional prefixes as spring or no spring, using criteria fixed beforehand and without later history. Their comparison table has eight joint spring labels, one disagreement in each direction and zero joint no-spring labels.

Observed agreement is 8/10 = 80%. Each rater uses spring nine times. Independent labels with those same marginal frequencies give expected agreement 0.9 × 0.9 + 0.1 × 0.1 = 82%. Defining kappa as (observed − expected)/(1 − expected) gives (0.80 − 0.82)/0.18 = −1/9, about −0.11.

For comparison, a separate balanced table 4,1 / 1,4 also has 80% agreement, but each rater uses each label five times. Expected agreement is 50%, giving kappa 0.6. Identical agreement percentages need not describe identical marginal label patterns. Neither calculation is a profit probability, a truth judgement or a measured real-market reliability study.

Keep interpretation separate from a complete trading test

Retain unresolved, rejected and inconvenient cases. Examine disagreements rather than treating 80% as certification. Ten assigned cases are especially sensitive to single changes, and selecting attractive completed drawings cannot evaluate recognition available at the time.

The next operational Wyckoff example connects frozen rules and decision timing. Related SMC terminology does not make institutions visible in candles. Record fills, costs, sizing and exits separately; neither a fifty-chart quota nor an eight-to-eighteen-month schedule certifies an edge or readiness for live trading.