Immediate definition
In plain words — For six weeks, a group applied Hurst's method to live trades and the book published unusually high results. The figures describe that sample; by themselves they do not prove repeatability.
The experiment ran from 28 October to 10 December 1968. Hurst presents it as a real-time application of the system developed in the preceding chapters. The team had five members and worked from a list of selected and analysed stocks, with one person continuously monitoring brokerage terminals.
Reported protocol
The initial objective was about 10% per trade with a mean interval of thirty days. To collect more cases within six weeks, the group arbitrarily set a profit-taking threshold at +11.1% gross. The book says signals were recorded as they occurred and communicated daily through memoranda to impartial observers; some trades were marked as actually funded.
This description is more informative than a chart selected after the fact, but it is not an independent replication. Cyclepedia does not possess the complete primary collection of memoranda, order instructions, and market data available at each moment. The accessible source remains the book written by the method's author.
Reported results
The account counts 42 transactions in 35 stocks. Thirty-eight are called successful and four unsuccessful: 38/42 equals 90.5%. The reported mean gross return is 11.1%; subtracting 2.2% mean costs gives 8.9% net, not 8.8%. The mean interval is 9.7 days.
The profit threshold directly shapes these figures: it deliberately closes many positions near the same return and shortens the observed duration. Moreover, 42 trades and 35 stocks are not necessarily independent observations. Market regime, group signals, and repeated positions may share the same underlying information.
Annualisation explained
Compounding 8.9% every 9.7 days produces an end value around 24.7 times the initial capital in 365 days: roughly a 2,374% gain, or 2,474% of starting capital. The arithmetic is coherent; the inference is not automatic. It assumes frequency, capacity, costs, and success remain unchanged as capital grows and regimes change.
The theoretical path from $10,000 to $1 million also takes about 54 intervals: 54 × 9.7 days is roughly 524 days, close to 17 months, not 15. Neither figure represents a year or seventeen months actually observed in the six-week sample.
What would be required today
A modern evaluation would define rules and universe before testing, include every signal, use point-in-time data, and account for spread, slippage, non-fills, and capacity. It would separate development from an out-of-sample period, release its logs, and disclose how many method variants were tried.
These controls matter because selecting the best configuration from many
attempts can create impressive apparent performance that degrades out of
sample. The 1968 case remains detailed historical testimony about the method's
use, but not independent validation. Source-checked means the figures and
caveats were checked against the source; it does not mean Cyclepedia confirms
the return.
Sources
- J. M. Hurst, The Profit Magic of Stock Transaction Timing, Prentice-Hall, 1970, Chapter 8, pp. 135–140 and Figures VIII-6–VIII-9 — primary source for the account.
- David H. Bailey, Jonathan M. Borwein, Marcos López de Prado, and Qiji Jim Zhu, Pseudo-Mathematics and Financial Charlatanism, 2014 — effects of backtest overfitting; it does not analyse Hurst's experiment.
- Campbell R. Harvey, Yan Liu, and Heqing Zhu, … and the Cross-Section of Expected Returns, NBER Working Paper 20592, 2014 — multiple testing and evidentiary burden.
Links
- Trading by logic — the integrated system.
- Screw and Bolt case — the single anatomised trade.
- Compounding and trading interval — the arithmetic behind annualisation.