Skip to main content
Track Head and shoulders
35 / 37
Search the archiveLibrary
Library

2013issue C0526-35

Algorithmic head-and-shoulders construction

A completed closing-price head-and-shoulders is outlined only after closing price crosses the neckline, using named milestone-points. Nine geometric tests of height, slope, timing, and intervening swings then decide whether that outline is accepted. The identification-algorithm is a recognition procedure, not a trading system.

  • Outline a completed closing-price head-and-shoulders only after the close crosses the neckline, using named milestone-points.
  • Apply nine geometric tests so shoulder height, neckline slope, flank timing, and intervening swings can each reject the outline.
  • Keep dist1 and dist2, plus the outer-trough flanks, inside fixed multiples so the peaks and troughs stay temporally balanced.
  • Treat the identification-algorithm as pattern recognition. An interpolating sketch and 0.80-to-1.20 scale factors can search longer or less symmetric cases.
Entries in this reading1 entry
Current trackHead and shoulders64 readings

Milestone-points after the neckline

A completed closing-price head-and-shoulders is outlined only after closing price has crossed the neckline, using a sequence of named milestone-points. Those points are the highs and lows that mark the shoulders, the head, the two interior troughs on the neckline, and the outer troughs.

The neckline is the line through the two interior troughs. Pattern completion is defined when closing price crosses that line. The head is the highest peak between the two shoulder peaks in the closing-price structure.

Shoulder placement

The shoulders are the left and right peaks that must stay high enough above their adjacent troughs and far enough below the head. The left-shoulder peak must sit above a 20 percent lift from its trough toward the head and at least 15 percent of that same head-to-trough span below the head.

The right-shoulder peak must sit above a 25 percent lift from its trough and at least 25 percent of that span below the head.

Interior troughs and the neckline

The right interior trough must stay within 15 percent below and 40 percent above the left interior trough, measured on the left-head vertical span. Extremely downward-sloping necklines are excluded by that band.

If the right interior trough is more than 20 percent of the left-head span above the left trough, one extra swing check applies. The highest close in the rightmost third of the head-to-right-trough interval must stay below 70 percent of the remaining rise from that trough to the head.

Relative shoulder height

Shoulder heights must differ by less than the shorter shoulder. Each shoulder must exceed 25 percent of its head-to-trough span unless the shorter of the two peak-to-adjacent-trough times is more than 25 percent of the trough-to-trough bar count.

Flank timing

Peak-to-peak timing is acceptable only when the longer of the two head-flank distances is under 2.5 times the shorter. Those flanks are dist1, the bar count from the left-shoulder peak to the head, and dist2, the bar count from the head to the right-shoulder peak. The longer of the two outer-trough flanks must stay under three times the shorter.

Intervening swings

Between the left trough and the head, the gap between the highest close in the left half and the lowest close in the right half must stay under half the head-to-left-trough height. That bound keeps the head from dissolving into a large intervening swing.

On the right shoulder, the first-third high must not sit above 80 percent of that shoulder's height unless the last-third low stays at least 40 percent of that height above the right trough.

Bounds before and after the pattern

The post-pattern high after the right outer trough must stay below one-third of the rise from that trough to the right-shoulder peak. The pre-pattern high before the left-shoulder peak must stay at least 15 percent of the left-head span below the head.

An identification-algorithm, not a trading system

The nine geometric tests plus the prior milestone-points scan form an identification-algorithm: a loop-light checklist of vertical, temporal, and swing tests that flags a head-and-shoulders outline from those points. The procedure is a pattern-recognition method, not a trading system.

A simple interpolating function can sketch the outline. Integer duration and 0.80-to-1.20 scale factors can search longer or less symmetric cases.

Eight Boyd Gaming H&S durations from the recursive scan

A trader should see that the recursive identification scan flagged eight completed closing-price head-and-shoulders on Boyd Gaming, and that only the short February 2010 case failed to precede a further drop. Bar heights are the formation lengths in bars, taken from the article’s table for this figure rather than estimated from the price pane.
A trader should see that the recursive identification scan flagged eight completed closing-price head-and-shoulders on Boyd Gaming, and that only the short February 2010 case failed to precede a further drop. Bar heights are the formation lengths in bars, taken from the article’s table for this figure rather than estimated from the price pane.Boyd Gaming (BYD) · daily · 2008-02-15T00:00:00.000Z to 2011-10-02T00:00:00.000Z

Siligardos warns that seven-of-eight declines on this sample overstate the pattern’s standalone power; extra filters and measured targets are still required before taking the short side.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
35 of 37 in the Head and shoulders track
201818-21 pp.Next on Head and shouldersInternational relative strength as a double-top case studyThe historical workflow identified a double top on the EFA-versus-U.S. large-cap relative-strength line at the May 2008 and September 2009 peaks, then described a long decline through March 2017.
All readings on this track · 37 readings
  1. 1982Head and shoulders as a three-path completion test
  2. 1984Stock low clusters as a cycle baseline
  3. 1985Four-phase construction of the head-and-shoulders reversal
  4. 1989Volume-confirmed reversal patterns, stops, and measured objectives
  5. 1991The journal as one checklist for taken and skipped trades
  6. 1991Head and shoulders as a direction hypothesis
  7. 1991Candlestick body and shadow construction with three-Buddha peaks
  8. 1992A three-count drill that binds candlesticks, head and shoulders, and entry rules
  9. 1997Constructing bump and run reversal channels
  10. 1998Testing reversal formations in bond futures
  11. 1999Construction first: extra shoulders, the neckline, and the diamond test
  12. 1999Dead-cat bounce, rollover, and failed reversals
  13. 1999Evaluating time gaps in bond reversal patterns
  14. 2000Constructing head and shoulders and double reversal patterns
  15. 2001Constructing broadening and complex bottoms
  16. 2001Constructing a slanted head-and-shoulders when the chart is tilted
  17. 2002Head and shoulders with dominant-cycle timing
  18. 2002Trendline breaks, right shoulders, and trailing stops
  19. 2003Confirmation tests for bearish top patterns
  20. 2003Commodity top hypotheses on a dollar rebound
  21. 2003A head-and-shoulders test during a bear rally
  22. 2004Pattern breakouts need a primary-trend filter
  23. 2004Reading candlestick closes on trendline and neckline tests
  24. 2004Candle diagnosis needs Western targets and stops
  25. 2004Head-and-shoulders neckline construction
  26. 2005A familiar chart condition is a hypothesis, not a completed decision
  27. 2005A 50-day ceiling and a rising-floor stalemate
  28. 2006A complete trading plan from philosophy to checklist
  29. 2006Thin-market head and shoulders with two averages and MACD confirmation
  30. 2010Head and shoulders as a playback-tested setup
  31. 2011Turning a head-and-shoulders outline into a breakout hypothesis
  32. 2011Volume-confirmed head and shoulders on AIG and Citigroup in 2007
  33. 2013Constructing head-and-shoulders milestone points
  34. 2013Head-and-shoulders geometry versus the filter stack
  35. 2013Algorithmic head-and-shoulders construction
  36. 2018International relative strength as a double-top case study
  37. 2019Structure invalidation before comfort-stops
By fieldChart patterns222 readingsBy year201398 readings
All 64 readings tagged Head and shoulders
Also on Head and shoulders5 readings
1982Head and shoulders as a three-path completion test1984Stock low clusters as a cycle baseline1985Four-phase construction of the head-and-shoulders reversal1989Volume-confirmed reversal patterns, stops, and measured objectives1990Constructing continuation patterns from congestion geometry