Immediate definition
In plain words — Even a coherent analysis can fail: news may arrive, the oscillation may change, or the chart may conceal a slower movement. Chapter 10 turns these cases into invalidation signals.
Hurst does not present his method as immune to the unexpected. He separates three families of failure and assigns an operating response to each. The map is historical, but its main lesson remains useful: before opening a position, define what would make the thesis false and how damage would be limited.
1. Security-specific shock
A sudden fundamental event may alter demand or supply before the cyclic model can adapt. Hurst separates a favourable shock, which invokes terminal-exit criteria, from an adverse one, left to the trailing loss level. The useful conclusion is not that a stop makes risk certain. It is that information outside the model can invalidate the thesis.
In actual execution, a stop is only a trigger. It becomes a market order and may fill far from its level during a gap or fast move; a stop-limit controls price but may remain unfilled. Size, liquidity, and an emergency scenario must therefore be planned together with the level.
2. Changing amplitude or duration
The second pitfall is magnitude-duration fluctuation: an oscillation may fade or change duration. In the book's language, constant-width envelopes that are no longer filled, and formations such as triangles, coils, wedges, or diamonds, warn that behaviour is changing.
When this occurs in a component shorter than the trading cycle, Hurst tries to use it as timing information. When it affects the trading cycle, he recommends temporarily avoiding the stock. A modern reading must prevent retrospective diagnosis: the same compression may have different outcomes. Exclusion rules must be defined beforehand and checked across all cases.
3. A horizon that is too short
A limited window may make a short component look dominant while a slower move changes direction. Hurst recommends more history, computational checks, and a different stock if doubt remains. More data is not an automatic cure: it also creates more lows, durations, and parameterisations from which to choose.
Risk control therefore needs a declared window, settings consistent across securities, and a record of rejected alternatives. Changing the horizon until the chart “works” introduces selection after the event.
An invalidation procedure
Before entry, record the data source, analysis version, invalidating event, intended order, and maximum acceptable loss under a worse fill. During the trade, distinguish model error, anomalous data, and execution trouble. After exit, retain the negative case too: without failures, the actual frequency of these pitfalls cannot be estimated.
The source-checked level certifies fidelity to the three cases in Chapter 10.
It does not certify that Hurst's remedies remove risk or that the chart
formations have current predictive effectiveness.
Sources
- J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970, Chapter 10, pp. 158–160 — primary source for the three pitfalls and historical remedies.
- SEC Office of Investor Education and Advocacy, Stop, Stop-Limit, and Trailing Stop Orders — price and non-execution risk.
- Campbell R. Harvey, Yan Liu, and Heqing Zhu, … and the Cross-Section of Expected Returns, NBER Working Paper 20592, 2014 — multiple testing and configuration selection.
Links
- Psychological barriers — the other half of Chapter 10.
- Cyclic cut-loss — the historical risk response.
- Half-span and full-span — Hurst's computational check.