CYCLEPEDIAThe trading encyclopedia
IT
James Marsden Hurst 1924—2005

Chapter 2.3 — Timing Is the Key

Hurst nominal cycles

The book's Table II-1 (1970): nominal market-cycle durations from 18 years down to 1.625 weeks, and how to use them to measure real cycles.

On this page

Who this entry is for — Hurst's "table of clocks": the reference durations against which every cycle measured on a chart is compared. Without this compass, every analyst counts weeks his own way.

Source: J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970 — Chapter 2, Table II-1 and The Nominality Principle (p. 33).

Prerequisites

Five principles of the cyclic model — in particular what "nominal" means and why real values vary.

Definition

In plain words — These are not periods fixed to the day. They are reference labels (13 weeks, 26, 18 months…). On the chart you measure the distance between two lows and see which row of the table you are approaching.

The nominal cycles are the reference durations imposed by the nominality principle: variation forces the model to be quantified with common values, from which every issue and every era deviate within a range. Nominal durations are an element of commonality; the deviations are the expression of variation.

Table II-1 (Hurst 1970, p. 33)

The original table has three columns — the same duration expressed in years, months and weeks where meaningful:

Years Months Weeks
18 — —
9 — —
4.5 — —
3 — —
1.5 18 —
1 12 —
0.75 9 —
0.5 * 6 26
0.25 * 3 13
— 1.5 6.5
— 0.75 3.25
— 0.375 1.625

* The table's original footnote: the 26- and 13-week components often appear in data as a combined effect of ~18-week nominal duration — which is exactly what the book measures on Warner Co. (18.5 weeks) and Standard Packaging (18.75).

HURST 1970 · CH. 2 The nominal cycle ladder Table II-1 in visual form · log horizontal axis, thickness ∝ magnitude CYCLEPEDIA DIAGRAM — EMICICLO 18 years 939 wk 9 years 470 wk 4.5 years 235 wk 3 years 157 wk 18 months 78 wk 12 months 52 wk 9 months 39 wk 26 weeks* 26 wk 13 weeks* 13 wk 6.5 weeks 6.50 wk 3.25 weeks 3.25 wk 1.625 weeks 1.63 wk * The 26- and 13-week components often appear in data as a combined ~18-week effect. Twelve reference clocks, from eighteen years down to eleven days.
The table in visual form: each duration is roughly half (sometimes a third) of the one before; longer bars are also thicker — proportionality.
Tap the info point

How to measure: the book's exercise

In plain words — Envelope on the chart, letters where price touches the bounds, count the weeks between lows. Average the good samples → current duration. Spread → expected variation.

Chapter 2 runs the full measurement on the weekly DJIA 1965–69 (Fig. II-3). The distances between the lettered lows:

Span Weeks Span Weeks
A–B 23 F–G 22
B–C 14 G–H 24
C–D 9 H–I 17
D–E 21 I–J 20
E–F 12 J–K 23

Ten samples; B–C, C–D and E–F are obvious variants (magnitude-duration fluctuation at work) and are discarded. The average of the remaining seven is 21.428 weeks: the current expression, in the DJIA, of the nominal 26-week component. The deviations (+2.572 / −4.428) round to ± 3.5: the near-future expectation is "21.4 ± 3.5-week cycles" — to be updated continuously, because variation never sleeps.

The chapter's measurements, all together

Nominal component Empirical measure Data
26 weeks 21.4 ± 3.5 (7 samples) DJIA weekly 1965–69, envelope
18 months (78 wk) 67 and 75 → 71 ± 4 same chart, nesting up
6.5 weeks 21.4 ÷ 3 = 7.14 → refined to 6.766 (15 samples, ±0.8) nesting down, Figs. II-5/II-6
4.5 years (54 months) 52 ± 1 months (4 samples) DJIA monthly 1949–69, log scale — the infamous "bull-bear" cycle (Fig. II-8)
13+26 → ~18 weeks 18.5 (Warner Co.) · 18.75 (Standard Packaging) numerical analysis and envelopes
Fig. II-5 — straight-segment envelope between lows
The original plate: the straight-segment envelope between weekly lows and highs — the technique that refines the estimate to 6.766 weeks.
Fig. II-7 — monthly DJIA on log scale
The original plate: monthly DJIA from 1949 on a log scale — four samples of 52 ± 1 months ≈ the 4.5-year nominal.

Example — You count 22 weeks between two lows? You are on the 26-week row, deviation −4: inside (or nearly) the ±3.5 range the book measured on the Dow. You count 19? Mind the table's footnote: you may be looking at the combined 13+26 effect.

Operating use

  1. Build the envelope on the chart.
  2. Count the weeks between consecutive lows (lows are better defined than highs, Hurst warns).
  3. Discard the obvious variants, average the rest, record the deviation.
  4. Compare with the nearest row of the table — and update at every new low.
  5. You need at least 6–7 data points per cycle: if you cannot fit them, expand the scale (daily); if you run out of samples, contract (monthly).

Summary card

If you measure… Nominal row Note
~5–8 weeks 6.5 wk the "gallop" inside the trading cycle
~11–15 weeks 13 wk
~17–20 weeks (13+26 combined) the ~18-week effect of the footnote
~20–28 weeks 26 wk the book's trading-cycle reference
~60–80 weeks 18 months measured by nesting up
~4–5 years 4.5 years the "bull-bear" cycle; monthly data required

Intraday extension (post-book tradition)

The nominal model of the Cycles Course (~1973) and the later tradition extends the ladder below the week, down to cycles of minutes, keeping the near-harmonic structure. It is not tabulated in the 1970 book: see Eight principles of the cyclic model and After the book.

Suggest an edit My proposals