CYCLEPEDIAThe trading encyclopedia
IT
James Marsden Hurst 1924—2005

Chapter 8.4 — Trading by Logic

The 1968 trading experiment

The book's field test: 28 October – 10 December 1968, a team of five, 42 transactions in 35 issues, 90.5% success, +11.1% average gross in 9.7 days — with memoranda to impartial observers.

On this page

Who this entry is for — The numbers trumpeted on the book's front page were born here: six weeks of real-time operations, documented day by day. This entry reports them all — and re-computes them, including the spots where the book's enthusiasm outruns the arithmetic.

Source: J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970 — Chapter 8, A Trading Experiment → Conclusions (pp. 135–140, Figs. VIII-6/VIII-9).

Prerequisites

The whole method: the experiment used every technique of Chapters 2–7 at "the absolute minimum" of analytical effort.

The experimental design

Card — The rules (28 October – 10 December 1968)

  • Objective: 10% per trade at a 30-day average interval.
  • Profits taken at an arbitrary +11.1% — deliberately, to hold the interval down and build the largest possible sample of trades.
  • Team of five: weekly scan, selection, analysis, tracking; one member at brokerage displays for the signals.
  • Documentation: every signal recorded as it occurred and reported by daily memorandum to impartial observers; asterisks = funds actually invested.

The results

HURST 1970 · CH. 8 The 1968 experiment: the 35 trades Fig. VIII-9 as data: each bar is a real trade (28 Oct – 10 Dec 1968) CYCLEPEDIA DIAGRAM — EMICICLO 5% 10% 15% mean +11.1% RESULT +11.1% gross / 9.7 days NET (2.2% COSTS) 8.9% → ≈2,474%/yr SUCCESSES 38/42 = 90,5% Not a backtest: daily memoranda to impartial observers, real funds on the asterisks.
Fig. VIII-9 turned into data: each bar a real trade, the gold line the mean (+11.1%).
Tap the best and the worst
Metric Value
Completed transactions 42 in 35 issues
Successes 38 (4 failures) = 90.5%
Average gross return +11.1% per trade
Average costs 2.2% → 8.9% net
Average interval 9.7 days (1.14%/day)
Equivalent annual yield ≈2,474%/yr (against the 313.8% objective)
Extras 3 market turns predicted (Fig. VIII-6) + 2 industry-group turns (average gains 21.5% and 21.2% — safety factor >2 on the objective)

The Dow sequence deserves mention: the dip to 930–950, the rise to 960–1000 and the steep drop toward 890–930 were all called in advance — with the predictions refined in flight (12 November, 25 November, 4 December) before the ~1000 top was even reached. The market followed the script even after the experiment closed.

Checking the arithmetic (editor's note) — 8.9% net every 9.7 days genuinely compounds to ≈2,474%/yr: 1.089^(365/9.7) = 24.7 ✓. But the front page's claim "$10,000 → $1,000,000 in approximately 15 months" is optimistic: it takes 54 trades at +8.9%, i.e. 54 × 9.7 = 524 days ≈ 17 months. And above all: six weeks with a team of five in a trending market are not "forever" — the book's own conclusions say as much.

The chapter's conclusions (honest, and the book's own)

  1. Intervals on the order of ten days are achievable, with signals sufficient for 100% time investment.
  2. The theoretical maximum of ~2,400%/yr (Appendix III!) can be produced.
  3. The 90% accuracy expectation was met — not exceeded.
  4. The average individual will do best concentrating on a few thoroughly analysed issues: handling many demands prohibitive time "unless a staff of personnel is utilized".
  5. High accuracy removes the need for wide diversification: best to limit commitments to 2–4 issues at a time.
  6. "The theory is valid, and the methods work. All you need to do is to master the techniques — and avoid the psychological barriers" (the bridge to Chapter 10).
Suggest an edit My proposals