In one sentence — When highs and lows draw closer, the chart forms a triangle; for Hurst the contraction may reflect wave amplitude, but it does not reveal the breakout direction by itself.
Shape first, hypothesis second
A triangle contains progressively narrower swings. Flags, pennants, coils, boxes, diamonds, and wedges are different names for pauses or contractions. The label helps describe what is visible; it does not automatically identify the process that produced it.
In Chapter 3 Hurst tries to bring this family into his composite-motion model. The question becomes not “which drawing promises a rise?” but “which components are shrinking, turning, or temporarily offsetting one another?”.
First mechanism: changing amplitude
Hurst uses magnitude-duration fluctuation for changes over time in the amplitude and duration attributed to components. If a short component loses amplitude while longer components change little over the observed interval, highs descend and lows rise: the oscillation fills less of its range and a triangle appears.
This is a construction inside the model. It does not mean every fall in volatility is a measurable cyclic component, or that amplitude must contract smoothly.
Second mechanism: a longer component turns
A similar shape may appear without an independent contraction in the short component. If a longer component reaches a high or low while shorter oscillations continue, the sum can produce converging peaks and troughs. This is the same geometric family that yields head and shoulders or double tops under a different timing relationship.
The two mechanisms are not interchangeable: one concerns mainly amplitude, the other the phase position of a longer component. Their final outlines may look alike, so the contour alone cannot select the explanation.
Where does breakout direction come from?
In Hurst’s historical reading, direction does not come from the word “triangle”. It comes from a separate estimate of component states: which are rising, which are falling, and how much of each turn has elapsed. If the components judged dominant agree, the model states a direction; if they disagree, uncertainty remains.
The 1961 Perkin‑Elmer case is the book’s example: Hurst compares two triangles with positions assigned to four periodicities. It is a detailed retrospective illustration, not a statistic covering all triangles.
Turning the claim into a modern test
A test needs boundaries set before observation: minimum number of swings, contraction rate, completion point, outcome direction and horizon, costs, and an out-of-period sample. The study by Lo, Mamaysky, and Wang shows one way to convert chart shapes into algorithmic objects; it does not validate Hurst’s particular cyclic causality.
Evidence boundary — The model explains how contraction can form. Shape alone does not establish breakout direction, probability, or profitability and may have several compatible explanations.
Sources
- J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970, Ch. 3, pp. 58–61, Figs. III-8 and III-9 — triangles, magnitude fluctuation, and the Perkin‑Elmer case.
- Andrew W. Lo, Harry Mamaysky, and Jiang Wang, Foundations of Technical Analysis, NBER Working Paper 7613, 2000 — algorithmic formalization and empirical testing of chart patterns.
Related entries
- Perkin‑Elmer case (1961) — historical application
- Double top and bottom — the other turning mechanism
- Chart pattern verification — the Chapter 3 framework