Who this entry is for — The classic trailing stop has a dilemma: too tight and a dip throws you out, too loose and you hand back the gain. Hurst resolves it by choosing which cycle to trail — and near the top he flips the buy method into a sell signal.
Source: J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970 — Chapter 5, Extension to Trailing "Sell" Signals and How to Construct Selling Analogs (pp. 88–91, Figs. V-1/V-2).
Prerequisites
The cyclic cut-loss — the first rung of the ladder climbed here.
Which cycle to trail
In plain words — You do not trail price: you trail the confirmed lows of the cycle next shorter than your trading cycle. A low is confirmed when price has already pulled away from it on the upside.
The arbitrary trailing stop fails like the arbitrary stop, and the cure is the same: levels dictated by the model. The book's rule, by stages:
- The initial stop is the first level.
- If several periodicities sit between the shortest component and the trading cycle, keep the original level until a low of the component next shorter than the trading cycle arrives; as prices pull away confirming it, that low is the new level.
- The trading cycle normally contains 2–3 of these cycles: at its first confirmed low your level already assures "a fair amount of profit retention". (Had you kept using the lows of the shortest cycle, you would probably have been sold out right there — the book makes the point deliberately.)
- Position within the envelope as compass: mid-channel, stay on the level from step 2; near the upper bound, switch to the shortest cycle's lows and, better still, to a series of uptrend lines.
On Gruen (19-week trading cycle, the 4-week as guide): levels TLL 1, 3, 4 are its successive lows; TLL 2, in between, is licensed by the violence of the pull-away from the first two-week low — with timing that good, you may validly "up" the level a little early.
Warning — As long as emergency levels and valid uptrend lines are not violated, the signal is "hold". You do not sell "because it went up a lot": you might be inside one of those sharp fundamental moves that override cyclicality upward — "and you'd hate to miss the ride". If the override is downward instead, the levels are there for exactly that.
The take-profit: the valid uptrend line
In plain words — Near the top, the buy method flips: uptrend lines on the rising lows, on ever shorter cycles, and the downside break of the "valid" one is the sell signal.
Exiting on trailing levels alone sacrifices "a goodly part of the maximum gain potential": you need the buy signal's analog. Every pull-away point from a trailing level is a point on an uptrend line; approaching the top you switch to lines built on ever shorter cycles — the series steepens, mirroring the downtrend lines of the buy zone.
On Gruen the sequence is documented day by day: in the eighth week of the 19-week cycle a well-defined triangle → state table (sum >19 hard up; 19-week 9 along, topping; 4-week 2½ along, hard down; 2- and 1-week bottoming) → upside resolution imminent but limited: guard up. The breakout allows steepening the line from a 4-week base to a 2-week base; a 4-week low is expected in ~1.5 weeks, probably lower: better out before. A stall at the line, one last push, and the break at 12⅜ gives the signal. Validation: from there Gruen declined — "though cyclically!" — to 9 over the next five weeks.
Summary card
| Phase | Level/tool |
|---|---|
| After purchase | Initial stop (cyclic cut-loss) |
| On the climb | Confirmed lows of the cycle next shorter than the trading cycle |
| Violent pull-away | An in-between level licensed (Gruen's TLL 2) |
| Near the upper bound | Shortest cycle's lows + series of uptrend lines |
| Profit taking | Break of the valid uptrend line |
| Always | The envelope as guide; "hold" while nothing is violated |
Links
- Cyclic cut-loss — the first rung
- Non-real-time envelope — the confirmation that needs no triangle
- Profit management — Chapter 5's framework
- Hurst tradition — chapter index