Skip to main content
Track Average Directional Index
43 / 56
Search the archiveLibrary
Library

2012issue C1167-71

Average Directional Index cluster filters for trend-start signals

A conventional directional-movement display uses one lookback for strength and direction. A cluster version plots three short lookbacks for the Average Directional Index, the plus directional indicator, and the minus directional indicator so congestion at an extreme, not a single reading, can precede a rule-based long, short, or stay-out decision.

  • Read each of the Average Directional Index, the plus directional indicator, and the minus directional indicator as a same-component trio of short lookbacks so cluster congestion, not a single-period reading, marks candidate trend starts and possible tops or bottoms.
  • Treat a complete high-side cluster as a candidate trend-end condition and a complete low-side cluster as a candidate trend-start condition, and do not treat a peak reading as interchangeable with a full cluster.
  • The Average Directional Index is a direction-neutral strength reading, so the plus and minus directional indicators must agree on initiation or exhaustion before a rule-based entry fires.
  • Encode a named cluster or crossing as a condition that stays off until the pattern is true, then turns on, so the same rule can be dropped into an exploration, a system test, or a chart strategy.
Entries in this reading3 entries
Current trackAverage Directional Index71 readingsTrend following89 readingsRule-based entry174 readings

Three lines where a conventional display uses one

A conventional directional-movement display uses one lookback for the Average Directional Index, the plus directional indicator, and the minus directional indicator. The cluster version splits each component into its own three-line study so the same strength and direction information can be read more granularly.

The featured procedure plots three Average Directional Index lines plus three plus directional indicator lines and three minus directional indicator lines. Cluster congestion across that same-component trio of short lookbacks, not a single-period reading, is used to mark trend starts and possible tops or bottoms. One implementation uses default lookbacks of three, four, and five bars for each three-line study.

High-side and low-side clusters are not peak readings

One implementation treats a complete high-side cluster as all three lines bunched above about 70 to 90 and a complete low-side cluster as all three bunched below about 30 to 20. Those extremes are used as candidate trend-end and trend-start conditions.

A peak reading is not treated as interchangeable with a full cluster. In the historical workflow, a tight low-side bunch can sit ahead of a trend start, while a later high-side peak without that same tight cluster is read as a different state.

Strength is not a substitute for direction

The Average Directional Index line is a direction-neutral strength reading. The plus directional indicator and the minus directional indicator carry both strength and direction and therefore must be read together rather than in isolation.

Daily Dow 3-period ADX against the 30 and 90 cluster rails

On the daily Dow, the fastest ADX lookback presses the 90 rail through the May–June selloff, then sits on the 30 rail in mid-August before turning up with the September advance. Treat those bunches at the rails as the filter, not a single ADX print. Levels come from the TradeStation TAC_3x_ADX pane on $INDU; the last 73.61 point is the platform’s printed ADX(3), and earlier points are read off the plotted curve.
On the daily Dow, the fastest ADX lookback presses the 90 rail through the May–June selloff, then sits on the 30 rail in mid-August before turning up with the September advance. Treat those bunches at the rails as the filter, not a single ADX print. Levels come from the TradeStation TAC_3x_ADX pane on $INDU; the last 73.61 point is the platform’s printed ADX(3), and earlier points are read off the plotted curve.Dow Jones Industrial Average ($INDU) · Daily · 2012-03-20T00:00:00.000Z to 2012-09-13T00:00:00.000Z

The source pane plots ADX lookbacks of 3, 4 and 5 with horizontal rails at 30 and 90. Only the leading 3-period line is digitized; the slower traces overlap too tightly on the raster to separate cleanly.

From a named cluster to a testable rule

The same idea can be encoded as conditions that stay at zero until a named cluster or crossing pattern is true, then flip to one so the pattern can be dropped into an exploration, a system test, or a chart strategy.

A further platform variant generates signals from convergences and crossings of the three-line Average Directional Index, plus directional indicator, and minus directional indicator clusters above 90 and below 10.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
43 of 56 in the Average Directional Index track
201213-21 pp.Next on Average Directional IndexConfirming a trend start or turn with a triple ADX clusterA 14-period ADX above 20 and rising is treated as a trending market, while +DI and -DI supply the up or down direction.
All readings on this track · 56 readings
  1. 1986Cycle-aligned directional trend indicator
  2. 1987What crossover and directional entry rules actually compare
  3. 1988A directional-line cross needs a trend filter, an extreme-point rule, and a dollar stop
  4. 1988Constructing true range by offset addressing
  5. 1988Constructing directional movement from bar range
  6. 1988Average directional index construction: recursive smoothing and lookback offset
  7. 1988Staged Average Directional Index construction with Relative Strength Index confirmation and stop alerts
  8. 1988Average Directional Index construction with frozen true range and directional rules
  9. 1991Constructing the average directional index from range expansion and true range
  10. 1991Constructing five-session forecasts from stochastic, ADX, and MACD inputs
  11. 1993Constructing the average directional index from directional movement and true range
  12. 1993Confirming n-bar breakouts with ADX and DX filters
  13. 1994Constructing a Bollinger band-width trend filter
  14. 1994A pre-trade checklist that can refuse a long three ways
  15. 1997An ADX threshold and a moving average as a trend filter
  16. 1998Regime filters for mutated indicators
  17. 1999Building the average directional index from range extension and true range
  18. 2000Evaluating ADX, RSI, and moving averages in a multi-stock warehouse
  19. 2000Stochastic pop as a filtered continuation setup
  20. 2000Onset and exit from one average directional index
  21. 2002Joint ADX and MACD readout for trend strength and direction
  22. 2003Adaptive Donchian breakout with implied volatility and volume
  23. 2004The average directional index as a regime gate for the relative strength index and the stochastic oscillator
  24. 2004Constructing true-range-specified volume as a directional filter
  25. 2005Constructing a multi-filter penny stock breakout procedure
  26. 2005Construct one playbook that flips with session regime
  27. 2005Combining Bollinger Bands, the average directional index, and Fibonacci retracement on currency pairs
  28. 2006Assembling an adaptive price zone from double-smoothed averages
  29. 2006An ADX strength gate for MACD and the stochastic oscillator
  30. 2007Directional movement as a filter plus trigger
  31. 2007Constructing a veto-first trend permission stack
  32. 2007ADX gates for trend end, range, and reversal
  33. 2008Constructing a nine-cell directional-ratio grid
  34. 2008Average directional index and directional trend indicator lookbacks as trend-filter parameters
  35. 2008A holding-matched market lens from averages and directional-line crosses
  36. 2008A nine-cell directional scoreboard for multi-horizon entries
  37. 2010Building a Vortex Indicator from high-low distances
  38. 2010Constructing ADX, RSI, and MACD price filters
  39. 2011Constructing a volume zone oscillator with a moving-average and Average Directional Index regime filter
  40. 2011A volume zone oscillator conditioned by an Average Directional Index filter
  41. 2011Candlestick names need volume-price, ADX, and moving-average checks
  42. 2012Clustered average-directional-index traces as a trend-start filter
  43. 2012Average Directional Index cluster filters for trend-start signals
  44. 2012Confirming a trend start or turn with a triple ADX cluster
  45. 2013Constructing a late-entry stack from a signed DMI oscillator
  46. 2013A directional oscillator and its stochastic as a stacked timing filter
  47. 2013ADX cluster lookbacks are a locked specification, not a chart label
  48. 2013Combining moving averages, stochastics, and ADX in a daily scan
  49. 2015Assembling the Average Directional Index from directional movement
  50. 2016How an Average Directional Index filter and a breakout entry form one procedure
  51. 2016Score RSI and stochastic crossings only when ADX confirms the trend
  52. 2018Constructing an ADX filter for intraday breakouts
  53. 2018An ADX volatility gate for prior-day breakouts
  54. 2019Exponential deviation bands with a moving average, RSI and ADX
  55. 2020A normalized-slope trend filter from linear regression
  56. 2020Gating volatility-momentum divergences with a Trend filter
By fieldIndicators and filters682 readingsBy year201281 readings
All 71 readings tagged Average Directional Index
Also on Average Directional Index5 readings
2003Adaptive Donchian breakout with implied volatility and volume2007ADX gates for trend end, range, and reversal2007Constructing a veto-first trend permission stack1987What crossover and directional entry rules actually compare2000Onset and exit from one average directional index