CYCLEPEDIAThe trading encyclopedia
IT
James Marsden Hurst 1924—2005

Chapter 3.1 — Verify Your Chart Patterns

Chart pattern verification

Chapter 3 in one entry: the cyclic model explains why charting's patterns exist, repeat and work — and adds what charting lacks: the reference time and the exit direction.

On this page

Who this entry is for — This is Chapter 3's map. If the "X" model is valid, it must explain the patterns chartists have observed for decades. It explains them all — and in return, charting gives the model its alarm bells.

Source: J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970 — Chapter 3, Verify Your Chart Patterns (pp. 51–67).

Prerequisites

The price-motion model and the five principles. The chapter puts them to the test.

The chapter's double purpose

In plain words — First: give you confidence in the model, by showing it explains what chartists have always seen. Second: teach you to use patterns better than the chartist, by reading the state of the cycles inside them.

Before applying the model you must believe in it — and reconciliation with "the precepts of charting" is powerful evidence: decades of independent empirical observation agreeing with the theory. But the traffic is not one-way: once you understand why a pattern forms, the pattern becomes a tool for determining the state of the cycles — and from there the direction and timing of the exit.

The detail entries, one per family:

Family Mechanism Entry
Trend lines and channels One cycle + the sum of the longer = a channel; the break is a long cycle turning Trend lines and channels
Head and shoulders Long cycle rolling over beneath a short one; the neckline is a second trend line Head and shoulders
Double top/bottom, "V" Same cause, different phase of the short component Double top and bottom
Triangles, flags, wedges… Magnitude fluctuation inside a pause of the longer cycles Triangles
The real test bench Two triangles, four measured cycles, one predicted "failure" Case Perkin-Elmer
Moving averages Cycle filters; 10 and 30 weeks fall between the nominals; plot them −½ span Cyclic moving averages

Saucers, measured moves, gaps, islands, one-day reversals: for the book they are all identifiable and explainable the same way — and it is "a very good exercise for the reader" to prove it.

The problem charting cannot solve

In plain words — Which fluctuations do you use to set up the pattern? On the same chart, up and down trend lines coexist, patterns nest inside patterns. The chartist picks "by eye"; the model says: they all count, each at its own scale.

At any given time you can find valid trend lines in both directions, heads and shoulders inside the shoulders of a bigger one, triangles that are part of larger triangles. On what basis does the chartist attach significance to one and not the other? It is the "arty" part of the craft — and its missing link is cyclicality: there is no reference time period. The model provides it: all patterns-within-patterns are significant, and the magnitude-duration relationship quantifies each one's importance. Up/down/sideways state tables for every observable periodicity, plus the magnitude-fluctuation warnings (flags, unfilled envelopes): "a major step in reducing the art of charting to the science of prediction".

The chapter's eleven conclusions

  1. Patterns form and repeat because of the repetitive nature of cyclic price action.
  2. Trend lines and channels are a mandatory outgrowth of the model.
  3. Head and shoulders, double tops/bottoms and "V"s have the same cause: the time relationship between components decides.
  4. Triangles, diamonds, wedges, flags and pennants arise from magnitude-duration fluctuation (or, beware, from rollover mechanics: telling the two causes apart is essential).
  5. Cyclic analysis tells in advance which way prices will go when a pattern terminates.
  6. The model provides the reference times without which charting remains an art.
  7. Moving averages are smoothers.
  8. The span is the design variable that turns them into cyclic-visibility tools.
  9. The model explains the empirical 10- and 30-week values.
  10. An average becomes far more useful plotted properly (half a span back).
  11. Averages aid cyclic analysis: they establish the trend of the summed long cycles and identify unambiguously the highs and lows around which curvilinear channels should be built.

Warning — The chapter does not say "throw charting away": it says patterns and model work hand in hand. The pattern calls attention; the state of the cycles verifies or repudiates the expectation. Using only the former is betting; using both is reading.

Suggest an edit My proposals