Skip to main content
Track Adaptive moving average
3 / 24
Search the archiveLibrary
Library

1992issue C121-2

Constructing volatility-adaptive exponential smoothing

A variable-index dynamic average keeps the exponential-smoothing recursion and replaces the fixed mixing weight with a volatility ratio, so the same update can be rebuilt on a close, high, or low series.

  • A variable-index dynamic average is exponential smoothing whose mixing weight follows measured volatility instead of staying fixed.
  • The volatility ratio is a 30-day standard deviation of the series being smoothed, scaled by a reference standard deviation of 0.5.
  • The recursive update blends the latest observation with the previous average, scaled by that volatility ratio and a fixed coefficient of 0.065.
  • The same construction applies to daily highs or daily lows when that series is used in both the standard-deviation window and the update.
Entries in this reading2 entries
Current trackAdaptive moving average28 readingsExponential smoothing117 readings

Keep the recursion, change the weight

A variable-index dynamic average is built from exponential smoothing by letting the mixing weight change with measured volatility instead of staying fixed. Exponential smoothing is a recursive average that applies a constant mixing weight to the newest observation and the complementary weight to the previous average. An adaptive moving average is a recursive average whose mixing weight is allowed to change with a measured property of the input series. A variable-index dynamic average is an exponential smoother whose mixing weight is scaled by recent standard deviation relative to a reference standard deviation.

Form the volatility ratio on the series being smoothed

The volatility input is the ratio of a 30-day standard deviation of the chosen price series to a reference standard deviation. That 30-day standard deviation is taken over a 30-observation window of the same series being smoothed. The reference standard deviation used to scale the 30-day window is 0.5. A volatility ratio is the quotient of that short-window standard deviation and the chosen reference standard deviation.

Blend the latest observation with the previous average

The recursive update blends the latest observation with the previous average, with the blend scaled by the volatility ratio and a fixed coefficient of 0.065. The smoothing weight is the coefficient that determines how much of the latest observation enters the updated average. The volatility ratio lets that smoothing weight change with measured volatility while the recursion itself stays the same.

VIDYA on the closing-price worksheet

Closing prices from mid-August through 1 October 1992, with the five VIDYA prints the worksheet actually computes once the 30-day standard deviation is available. Traders should notice the adaptive average sitting near 105.5 while the close swings from about 103.2 to 106.9; the last five rows are taken from the source spreadsheet, not redrawn from a curve.
Closing prices from mid-August through 1 October 1992, with the five VIDYA prints the worksheet actually computes once the 30-day standard deviation is available. Traders should notice the adaptive average sitting near 105.5 while the close swings from about 103.2 to 106.9; the last five rows are taken from the source spreadsheet, not redrawn from a curve.daily · 1992-08-14T00:00:00.000Z to 1992-10-01T00:00:00.000Z

VIDYA starts only at row 31 because the 30-day STDEVP window needs D6:D35. The source fixes the reference deviation at 0.5 and the base alpha at 0.065 in =(E35/0.5*0.065*D35)+((1-(E35/0.5*0.065))*F34).

Rebuild the same update on highs or lows

The same construction can be applied to daily highs or daily lows by substituting that series into both the standard-deviation window and the recursive update. The window and the average stay on one series: a high-based smoother uses highs for the 30-observation standard deviation and for the recursive blend, and a low-based smoother does the same with lows. The historical presentation treats the method as an extension of the exponential moving average.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
3 of 24 in the Adaptive moving average track
19951-3 pp.Next on Adaptive moving averageConstructing an adaptive moving average with an efficiency ratio and filterTwo n-period inputs, net change as direction and the sum of absolute one-period changes as volatility, form an efficiency ratio that equals 1 in a one-way window and approaches 0 when net travel is small relative to the path.
All readings on this track · 24 readings
  1. 1991Building variable-length moving averages from partitioned price changes
  2. 1991Variable-length moving average from change dispersion
  3. 1992Constructing volatility-adaptive exponential smoothing
  4. 1995Constructing an adaptive moving average with an efficiency ratio and filter
  5. 1995Two-gate breakout confirmation with adaptive averages
  6. 1995Building momentum-scaled adaptive moving averages
  7. 1995Adaptive length as a construction choice inside exponential smoothing
  8. 1998Testing price-channel breakouts with a lag-aware adaptive average
  9. 1998Constructing filters by nesting offsets and variable weights
  10. 1998Constructing an efficiency ratio adaptive average and entry filter
  11. 2001Encoding candle structure as a numeric filter
  12. 2001Adaptive averages driven by cycle-phase speed
  13. 2005Constructing an adaptive moving average from a fractal-dimension weight
  14. 2005Range-dimension adaptive exponential filter
  15. 2010Constructing simple, exponential, and adaptive averages
  16. 2010How a price-hugging smoother is assembled from ordinary averages
  17. 2013Evaluating an adaptive moving average against a same-window moving average
  18. 2016Three-layer confirmation: adaptive average, stochastic relative strength index, and stop-and-reverse
  19. 2017One-alpha reverse-path exponential smoothing
  20. 2018Pair two adaptive averages to filter swing turns
  21. 2018Two Adaptive moving averages as a confirmation pair
  22. 2018Constructing an adaptive filter for adoption-cycle reversals
  23. 2018Constructing a deviation-scaled adaptive moving average
  24. 2020Walk-forward and adaptive averages as two tests of the same trend
By fieldIndicators and filters682 readingsBy year1992115 readings
All 28 readings tagged Adaptive moving average
Also on Adaptive moving average5 readings
1995Adaptive length as a construction choice inside exponential smoothing1998Constructing filters by nesting offsets and variable weights2005Constructing a fractal-dimension adaptive moving average2005Range-dimension adaptive exponential filter2010Constructing simple, exponential, and adaptive averages