Who this entry is for — Chapter 2 showed what the envelope is. This entry explains how Hurst drew it: the manual procedure, its validity over time and the state table used to turn channels into projections.
Source: J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970 — Chapter 4, How to Construct Curvilinear Envelopes and Prediction of Price Turns Using Envelopes (pp. 69–74).
Verification scope — The procedure and the statements attributed to Hurst were compared with the primary text. This check documents the author's method; it does not certify its predictive effectiveness.
Prerequisites
Curvilinear envelope and Hurst nominal cycles. This entry moves from the model to the manual charting procedure.
How much chart history is needed
In plain terms — The chart should show at least one and a half cycles of the component immediately longer than the one being traded, because its larger magnitude can dominate price movement.
The procedure starts with a weekly high-low chart covering at least 1.5 cycles of the periodicity immediately longer than the trading cycle and, where possible, the second longer component as well. The rationale is proportionality: longer cycles can move many more points than shorter ones and therefore can strongly influence price regardless of the current phase of the trading cycle.
The procedure, step by step
Card — Envelope construction (Chapter 4)
- 1. Identify the dominant component visually, usually a variation of the nominal 13- or 26-week component.
- 2. Lightly sketch the two bounds of a band following that movement; measure vertical thickness at several points, choose the mean and correct the band to constant width.
- 3. If observations remain outside, widen the band until all are included.
- 4. Repeat while narrowing. A minimum width will normally contain everything except two or three isolated points, recognisable because they lie alone between the wider and narrower bands. The wider band can then be erased.
- 5. Use contacts with the final band: mark well-defined lows, count weeks from low to low, exclude variants, calculate the mean and tabulate the result.
- 6. The highs and lows of the dominant component become points for a second envelope, drawn with the same technique. Two channels identify three components; a still shorter component is often visible within the narrow channel, but its lows can be marked without forcing a third channel.
The compromise between the two criteria — enclosing the data and minimising width — is explicit in the book. In the worked example, three brief overruns remain outside; Hurst explains that including them would require the channel to be widened considerably.
The cyclic state table
In plain terms — For each identified component, record its average duration and the date of its latest low, then estimate its current position. Applied to historical observations, the table organises the model's reading of earlier movements; applied to the latest observation, it becomes a projection.
Once average durations have been estimated, a table is prepared with the condition of each component before every major movement on the chart: the number of weeks elapsed since its latest low. In an ideal component, the peak lies halfway through the period; in a composite price series, the sum of the other components can shift it. Hurst argues that properly drawn channels allow the tabulations to explain the observed fluctuations. This is a claim of the model, not a universal empirical conclusion.
The table is then recalculated for the latest available observation. The setup sought by the method combines two or more components expected to reach lows at about the same time with an estimated upward contribution from the longer components. This leads to closer tracking and the signals described in Chapter 4.
Validity over time
Warning — In the graphical method, the portion extending backward from the latest well-defined high or low is treated as stabilised. The forward bound is an estimate that must be updated; Hurst indicates a maximum uncertainty on the order of half a cycle, reduced when a new extreme forms. Price outside the estimated channel is an anomaly to investigate, not unambiguous proof of a fundamental cause or of one specific analytical error.
The chapter adds two practical points. It is useful to check whether a whole number of shorter components, usually two or three, fits within the chosen trading cycle because each provides information about the next longer component. It is also useful to maintain a daily chart covering at least 1.5 trading cycles as a finer timing reference. If a security does not display sufficiently readable cyclic behaviour, Hurst's procedure is to select another one.
Software implementations
The book's procedure includes graphical steps and judgment in selecting contacts. Software can formalise them, but should disclose at least:
- the centre-line algorithm;
- the width rule and treatment of outliers;
- the criteria for contacts, troughs and excluded samples;
- the handling of the latest unconfirmed section;
- the difference between a calculation available in real time and an ex-post reconstruction.
A band based on a centred moving average is not automatically equivalent to the manual envelope. A theoretical article should not claim that a particular tool conforms to Hurst's method without a specification and reproducible tests.
Summary card
| Step | Rule |
|---|---|
| Data | At least 1.5 cycles of the component longer than the trading component |
| Width | Constant; the minimum that covers all but two or three isolated points |
| Second envelope | Drawn on the dominant component's highs and lows, yielding three identified components |
| State table | Weeks from the latest low; the idealised peak lies at half the duration |
| Setup | At least two lows expected together, with the estimated longer-component sum rising |
| Validity | Historical portion through the latest extreme; forward portion is an estimate with uncertainty on the order of up to half a cycle |
Sources
- J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970, Chapter 4, pp. 69–74 — envelope construction, width compromise and the latest section.
- J. M. Hurst, same work, Chapter 4, pp. 75–85 — cyclic state table, turning-point window and transition to graphical signals.
Links
- Graphic buy timing — how the state table is used
- Gruen Industries case — the procedure applied to blind data
- Curvilinear envelope · Nesting envelope
- Hurst tradition — chapter index