2012issue C1061
A two-stage case study in market cycle analysis
A usable market fast Fourier transform depended on a separate preprocessing stage, then the transform, then a comparison with another explicit Fourier baseline. A historical coding case shows how that workflow was split and later adapted as a user-written study.
- A usable market fast Fourier transform needed a dedicated preprocessing stage before the transform itself.
- One collaborator coded the transform and the other coded the preparatory treatment that had to run first.
- The finished routine was adapted as a user-written study so its cycle estimate could be compared with a Fourier study already bundled in the same software.
- Editorial reading: spectral analysis is this two-stage workflow, not the transform alone.
The case in brief
A practitioner who had already coded a fast Fourier transform in Fortran for a college engineering project later recoded that study in BASIC for personal-computer use.
A usable market version still required a dedicated preprocessing stage in addition to the transform calculation. The two parts were not interchangeable.
Preparation had to run first
Implementation was split. One collaborator, who held a mathematics doctorate, coded the transform. The other coded the data preprocessing that had to run first.
That preparatory treatment of the ordered series was treated as a separate stage. The fast Fourier transform was applied only after it was complete.
A user-written study and a bundled comparison
The finished routine was adapted to run as a user-written study in early personal-computer technical analysis software that accepted BASIC programs.
The authors judged their fast Fourier transform routine more accurate than a Fourier study already bundled with that software. They were invited to present the work at a user seminar.
Material developed for the Fourier implementation was reused in the first issue of a technical analysis periodical timed for distribution at that seminar.
Reading the comparison
Editorial note: the archive records a judgment against an explicit bundled Fourier baseline, not a present-day test. The useful method lesson is the order of work. Prepare the ordered price, volume, or breadth series, estimate frequency content over a stated sampling interval and lookback, then compare that cycle reading with another stated Fourier baseline.
All readings on this track · 16 readings
- 1982Building FFT spectra to size cycle filters
- 1988Fourier cycle models break in major swings
- 1988Constructing moving average filters from price Fast Fourier Transforms
- 1989Staging Fast Fourier construction under memory limits
- 1993Constructing forecast inputs with moving averages, Fourier transforms and intermarket spreads
- 1994Preprocessing prices so Fourier peaks set moving-average lengths
- 1994Constructing a spreadsheet FFT power spectrum from daily prices
- 1994Building dominant-cycle spectra with FFT preprocessing
- 1994Constructing labeled cycle lengths from FFT spectra
- 1999Fast Fourier Transform reconstruction is not a walk-forward decision tool
- 1999Walk-forward endpoint Fourier construction as a same-day mechanical procedure
- 2002From the power spectrum to indicator windows
- 2003Endpoint Fast Fourier Transform evaluation with walk-forward mechanical rules
- 2004Constructing signal and noise from market waveforms
- 2012A two-stage case study in market cycle analysis
- 2015Whitening pink noise to build a near-zero-lag cycle oscillator