Skip to content
CyclepediaThe trading encyclopedia
SESSION
ITENESPT-BRFRDE
James Marsden Hurst 1924—2005 Biography and works
Chapter 3.6 — Verify Your Chart Patterns

Perkin-Elmer 1961: Hurst's two-triangle case

Hurst rereads two triangles on Perkin-Elmer's 1961 daily chart through four estimated periodicities; it is an illustrative historical case, not a general test.

Cite this entry

This reference cites the Cyclepedia entry. The documents it draws on are listed in the entry’s sources.

On this page

In one sentence — In his 1970 book Hurst rereads a daily chart from 1961: two triangles, meaning contractions in price oscillations, end downward. He uses four estimated durations to interpret their outcome.

What is the document?

The case appears in Chapter 3 of The Profit Magic of Stock Transaction Timing. Hurst reproduces Perkin-Elmer's daily highs and lows from March to September 1961, taken, according to his text, from a popular charting manual (Fig. III-9, p. 60). Two triangles are marked: the manual classified the first as a failure of the continuation rule, while the second broke in the direction the chartist expected.

This reconstructs the reasoning Hurst applied to an already observed chart. The pages examined do not document a forecast recorded before the events of 1961.

The four periodicities attributed

Hurst counts intervals between lows, not just points marked on the chart. On p. 59 he reports these average durations and compares them with the nominal values, meaning the reference durations in his model:

Reported duration Count or relationship described by Hurst Nominal reference
13.9 weeks 2 intervals between lows A, B, and C 13 weeks
4.6 weeks 6 intervals between points 1 through 7 3.25 weeks
2.3 weeks 12 intervals, two within each 4.6-week interval 1.625 weeks
1.15 weeks Half the 2.3-week duration Shorter component, not tabulated there in the model

Three lows delimit two complete cycles; seven lows delimit six. Using the rounded averages, 13.9 / 4.6 ≈ 3.02, 4.6 / 2.3 = 2, and 2.3 / 1.15 = 2: the nesting described is roughly 3:2:2, not uniform halving. These are Hurst's counts on that chart, not guaranteed durations or a new independent estimate by Cyclepedia.

In the first episode he assigns the four components positions of about 8.0, 3.8, 1.3, and 1.1 weeks from the respective component's preceding low. His table judges all four to be falling (p. 60). These positions are elapsed times, distinct from the average duration of a whole cycle.

Perkin-Elmer 1961: four aligned estimates Interactive board reporting the durations, positions, and direction attributed by Hurst in his historical example. First break: Hurst’s reading Daily, March–September 1961 · positions reported in Chapter 3 Mean duration Since its own low Assigned state 13.9 weeks 8.0 weeks DOWN i 4.6 weeks 3.8 weeks DOWN i 2.3 weeks 1.3 weeks DOWN i 1.15 weeks 1.1 weeks DOWN i In the historical account: four directions down → break down Illustrative in-sample case, not a general success rate Cyclepedia diagram · Emiciclo
The durations and positions are those reported in Hurst’s historical account. Tap each row to read its role.
Select the highlighted points to explore the detail

First triangle: the decline in Hurst's reconstruction

In the author's reading, the small triangle contains three oscillations of the 1.15-week component with varying amplitude; Hurst describes it more precisely as a flag, a short consolidation formation. The chart rule he cites favored continuation of the preceding rise; Hurst's estimated cyclic state pointed down on all four scales. Price broke downward.

Hurst uses the episode to argue that a component-state table can anticipate when the chart rule will “fail”. The sequence agrees with his reading, but the parameters were derived and interpreted in the same case. This example therefore does not tell us the error rate on new charts.

Second triangle: same outcome, different construction

In the larger triangle Hurst sees one cycle and most of the next of the 4.6-week component (p. 60) while the 13.9-week component passes its high. At the chosen moment he assigns downward direction to the 13.9-, 4.6-, and 2.3-week components; the shortest is bottoming, flat to rising (p. 61). He also notes that the preceding 13.9-week low was lower, which he reads as weakness in longer components.

The break is downward again. Here charting and the cyclic reading agreed on direction, while Hurst argues that the latter supplied a mechanism.

What it teaches—and what it does not

The case helps distinguish measured durations, the state assigned to components, and the observed outcome. That is more informative than the label “triangle” alone. It does not establish that periodicities were stable, the decomposition unique, or the procedure effective out of sample after costs.

Lo, Mamaysky, and Wang study patterns on other data, without testing this episode or Hurst's model.

Evidence boundary — The 1961 chart documents how Hurst applied his method and an outcome consistent with his reading. It supplies no success frequency, out-of-sample comparison, or causal proof.

Sources

Suggest an edit My proposals