Skip to main content
Track Runs test
8 / 15
Search the archiveLibrary
Library

1992issue C071-5

Channel-height ratios for equity trend evaluation

Nested high-low price channels were turned into channel-height ratios and compared with a square-root baseline. The runs test asked whether expansion beat that no-trend forecast on a defined sampling interval and lookback.

  • Channel-height ratios were formed at lookbacks of 1, 4, 9, 16, 25, 36, and 49 sessions by dividing each average multi-session price-channel height by the average one-session range.
  • Trend tendency was scored only when a channel-height ratio exceeded the square-root baseline, the explicit no-trend envelope used in the runs test.
  • On the full sample, all seven equities sat above that envelope, the NASDAQ composite led the indexes at every lookback, and the Dow Jones Industrial Average stayed below the baseline at every lookback.
  • Editorial reading: a series that does not outgrow the square-root baseline across successive lookbacks fails the no-trend runs test and should not proceed to trend-following evaluation.
Entries in this reading3 entries
Current trackRuns test19 readingsPrice channel77 readingsTrend following89 readings

How the nested price channels were measured

A price channel is the high-low envelope taken over a fixed number of sessions. Its average height is the raw measurement used to score expansion.

Average high-low channel heights were computed at lookbacks of 1, 4, 9, 16, 25, 36, and 49 sessions. Each average was then divided by the one-session average to form a channel-height ratio, a dimensionless scale of multi-session height relative to the average one-session range.

The calculation window used daily observations from 2 January 1985 through 31 January 1992. Holiday rows were removed first, and only series with history before 1985 were admitted.

Scoring trend tendency with a runs test

Trend tendency was scored by asking whether the channel-height ratio exceeded the square root of channel length. That square-root baseline is the no-trend envelope for the evaluation, and a ratio above the envelope is treated as excess directional expansion.

The runs test is an ordered-range comparison that asks whether observed channel expansion beats an explicit random-walk forecast over a defined sampling interval and lookback.

Full-sample ranks among indexes and equities

On the full seven-year-and-one-month sample, the NASDAQ composite produced the largest index ratios at every reported lookback, while the Dow Jones Industrial Average stayed below the square-root baseline at every lookback.

Across the seven individual equities, every full-sample ratio sat above the square-root envelope. Eli Lilly posted the largest values and Xerox the smallest.

Index ratios generally increased with the number of constituents, yet the Major Market Index still outranked the Dow industrials on the same channel-height scale.

Alternative windows and a shorter cash index sample

Each series was recomputed on six windows: the full sample, the first year, the first two years, the last year, the last two years, and a middle year. For the OEX, window-to-window ratio gaps approached 10 percent, but the rise from short to long channels did not reverse.

The cash S&P 500 sample spanned only six years and seven months and was not interchangeable with a spliced-futures application of the same ratio method over a different interval.

Dow Jones Industrial Average, January 1985–January 1992

Over the seven-year-and-one-month window used for the channel-height study, the Dow advanced from the high-1200s to the mid-3300s, with the 1987 crash and the 1990 selloff as the two clear breaks. Levels were read from the published line chart against the 100-point grid, not from a numeric table, so they are approximate.
Over the seven-year-and-one-month window used for the channel-height study, the Dow advanced from the high-1200s to the mid-3300s, with the 1987 crash and the 1990 selloff as the two clear breaks. Levels were read from the published line chart against the 100-point grid, not from a numeric table, so they are approximate.DJIA · Daily · 1985-01-01T00:00:00.000Z to 1992-01-31T00:00:00.000Z

Digitized from the printed curve; readings are good to roughly 20–30 Dow points. The source study dropped holiday bars before computing channel heights; this series follows the published ink only.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
8 of 15 in the Runs test track
20011-7 pp.Next on Runs testA runs test before volatility and expected-value sizingMoney-management-indicators are computed from the last N closed trades and hold until a new exit, so they describe recent system behavior rather than the full trade history.
All readings on this track · 15 readings
  1. 1986Constructing runs and persistence tests from labeled prices
  2. 1986Evaluating daily price and volume serial independence windows
  3. 1986Evaluating advance-decline plus-day runs against chance baselines
  4. 1986Weekly resamples as a diagnostic filter for statistical windows
  5. 1988Runs test as a critique of price-series memory
  6. 1989Evaluating weekday close direction with a counted baseline
  7. 1989Statistical windows for indicator time parameters
  8. 1992Channel-height ratios for equity trend evaluation
  9. 2001A runs test before volatility and expected-value sizing
  10. 2005Constructing runs-test z-scores for signed return persistence
  11. 2005Evaluating persistence with runs and autocorrelation
  12. 2005Weekday FX turning points and close run tests
  13. 2013Constructing a runs-test turn forecast
  14. 2017Star rating from slope and swing runs
  15. 2018Regime-dependent odds after directional price runs
By fieldStatistical tests59 readingsBy year1992115 readings
All 19 readings tagged Runs test
Also on Runs test5 readings
1986Constructing runs and persistence tests from labeled prices1986Evaluating advance-decline plus-day runs against chance baselines1986Evaluating daily price and volume serial independence windows1986Weekly resamples as a diagnostic filter for statistical windows1987Evaluating money-supply serial dependence before a forecast