Skip to main content
Track On-balance volume
57 / 61
Search the archiveLibrary
Library

2019issue C0218-23

Volume-volatility parabola and signed-volume momentum

A 10-year SPY window ending September 29, 2017 is described as a parabola of 25-day mean return against 25-day mean volume, with a vertex volume of 148 million. The archive treats market momentum as the signed product of volume and price change, the construction brief used here for on-balance volume and volume-price analysis.

  • A 25-day mean-return versus 25-day mean-volume plot for SPY over a 10-year window ending September 29, 2017 is described as a parabola rather than a straight line, with a vertex volume of 148 million.
  • Below that vertex, 25-day mean return is described as oscillating near zero between -0.5% and +0.5%, and the series median volume of about 142 million sits close to the vertex.
  • In that same window, average volume on up days is about 4% below average and average volume on down days is 4% above average, while mean price change is +0.77% on up days versus -0.87% on down days.
  • The source treats market momentum as the signed product of volume and price change and states that an unfilled order has no momentum because no trade is completed.
Entries in this reading2 entries
Current trackOn-balance volume88 readingsVolume-price analysis207 readings

Mean return against mean volume

A 25-day mean-return versus 25-day mean-volume plot for SPY over a 10-year window ending September 29, 2017 is described as a parabola rather than a straight line, with a vertex volume of 148 million.

Below that vertex, 25-day mean return is described as oscillating near zero between -0.5% and +0.5%. The series median volume of about 142 million sits close to the vertex.

SPY 25-day mean volume versus 25-day mean return

On this 10-year SPY sample the fitted parabola bottoms near 148 million shares: quieter sessions keep the 25-day mean return inside about half a percent, and only the richer half of volume sits with large moves, more often to the downside. The coordinates come from the quadratic printed in the source for 25-day mean volume against 25-day mean return, evaluated along the published return axis, not from a hand-traced scatter.
On this 10-year SPY sample the fitted parabola bottoms near 148 million shares: quieter sessions keep the 25-day mean return inside about half a percent, and only the richer half of volume sits with large moves, more often to the downside. The coordinates come from the quadratic printed in the source for 25-day mean volume against 25-day mean return, evaluated along the published return axis, not from a hand-traced scatter.SPY · 25-day mean · 2007-09-29T00:00:00.000Z to 2017-09-29T00:00:00.000Z

Least-squares fit reported as y = 258,415,324x² − 102,208,800x + 158,194,769 on 25-day means for SPY through 29 September 2017 (R² = 0.3140). Volume is shown in millions of shares. The author notes the skew can flip with the window or the symbol; this is the printed parabola, not the underlying scatter.

Volume on gaining and declining days

In that same SPY window, average volume on up days is about 4% below average and average volume on down days is 4% above average. Mean price change is +0.77% on up days versus -0.87% on down days.

When volume is split by gaining versus declining days and scaled as a multiple of median volume, absolute price change rises with volume on both sides below about three times the median. Declining days move more than gaining days above that threshold.

A 10-year SPY window ending September 29, 2017 associates negative volume with a mean daily price decline of 0.15%.

Competing accounts of volume and volatility

The source frames four competing accounts of volume and volatility. Mixture of distributions treats information as near-simultaneous, with little lag. Sequential information arrival has informed traders act first. Dispersion of belief is a third account. Noise trading can keep a volume spike clustered for days.

Momentum only after a trade

The source treats market momentum as the signed product of volume and price change. An unfilled order has no momentum because no trade is completed.

Editorial reading for filter construction

TradersWeek reads the parabola as a reason not to treat volume as a yes-or-no stamp on price. Near the vertex, and near the nearby median of about 142 million, 25-day mean return sits close to zero, so a volume reading in that region is a weak construction input.

The signed product of volume and price change is the archive definition of momentum. That definition can replace the slogan that volume precedes price when on-balance volume or volume-price analysis is rebuilt as a filter.

The four accounts remain competing. Mixture of distributions implies little lag. Sequential information arrival has informed traders act first. Dispersion of belief and noise trading that clusters a volume spike for days are alternative constructions. The archive does not choose among them.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
57 of 61 in the On-balance volume track
201922-25 pp.Next on On-balance volumeSeparating on-balance volume into breadth and volumeOn-balance volume assigns a session's entire volume as positive or negative from whether the close is above or below the prior close, so the series responds to price direction rather than to the size of the price change.
All readings on this track · 61 readings
  1. 1988Cumulative volume trend and 200-session polarity
  2. 1988Three clocks when cumulative volume led price
  3. 1988Payoff index checkpoints for open interest and divergence
  4. 1988Relative strength, stochastics, and on-balance volume as a turning-point stack
  5. 1988Signing volume by open interest change
  6. 1988Case study: weekly on-balance volume with trendlines and a 21-period average
  7. 1988Combining weekly and daily MACD divergences with on-balance volume
  8. 1988Restoring the original tests for market breadth, on-balance volume, and trendlines
  9. 1989Evaluating volume-weighted breadth extremes
  10. 1989Demand oscillator construction with a momentum overlay
  11. 1989Constructing a volume percentage ratio from advance-decline flow
  12. 1990Constructing net positive volume as On-balance volume
  13. 1990Buying-share scale with on-balance volume
  14. 1990Friday-filtered on-balance confirmation lines
  15. 1991Constructing rate-market volume breadth from on-balance volume
  16. 1991Building a breadth composite from on-balance volume breakouts
  17. 1992Constructing a rate-complex volume-breadth index
  18. 1992Constructing a breadth index from on-balance volume
  19. 1992Constructing signed market thrust from issue volume
  20. 1992Constructing indicator phase across averages, volume and bands
  21. 1993Force on-balance volume to agree with itself at slower horizons
  22. 1993Three transforms of market thrust from breadth and volume
  23. 1994Constructing a tick-line momentum oscillator
  24. 1994Constructing a smoothed on-balance tick momentum oscillator
  25. 1994Histogram confirmation and a second-series veto for MACD-family alerts
  26. 1994Constructing a 40-period envelope on on-balance volume
  27. 1996Constructing a matched percent-B volume disparity
  28. 1997Constructing calendar regime overlays and volume-weighted rules
  29. 1997Constructing market thrust from breadth and volume
  30. 1997Constructing on-balance volume as a signed running total
  31. 1997A three-layer construction for a signed volume-force oscillator
  32. 1997Constructing a range-sum volume-force oscillator
  33. 1999On-balance volume confirmation, price-indicator divergence, and bounded exits
  34. 2000A daily long-only confirmation stack on the tether line
  35. 2000A three-filter tether and volume trend system
  36. 2003Constructing finite volume elements as a finite-period money-flow filter
  37. 2003Constructing on-balance volume from signed session volume
  38. 2003On-balance volume as a market-breadth participation ledger
  39. 2003Construct a volatility-adjusted on-balance volume filter before confirming a breakout
  40. 2004Opening and midday ranges as an on-balance-volume breakout case
  41. 2004Constructing a volatility-gated volume-flow measure
  42. 2004Constructing a volume-flow series from on-balance volume
  43. 2004Volume-flow indicator construction from on-balance volume
  44. 2004Volume-flow slope divergence, then separate on-balance-volume and crossover rules
  45. 2004Volume flow and band construction as one inspectable problem
  46. 2004Auditing advance-decline confirmation for scale, horizon, and composition
  47. 2005Bounded market breadth from the Arms index to BRIN
  48. 2006Overnight-hold entry and exit construction
  49. 2007Evaluating Volume-price confirmation versus On-balance volume and MACD
  50. 2009Volume confirmation versus a price advance
  51. 2010Typeahead retrieval of a moving average, on-balance volume, and a stochastic oscillator
  52. 2010Modified volume-price trend and log-price scale alignment
  53. 2010Constructing liquidity accelerometers from advance-decline averages
  54. 2012Failed short cycles as breadth and volume tests
  55. 2016Earnings straddles, implied volatility, and money-flow filters
  56. 2017A volume residual filter and a price projection matrix
  57. 2019Volume-volatility parabola and signed-volume momentum
  58. 2019Separating on-balance volume into breadth and volume
  59. 2020Constructing the obvm signal line
  60. 2020Constructing smoothed on-balance volume with a signal line
  61. 2020Keep neutrality as a first-class input in three-state market data
By fieldIndicators and filters874 readingsBy year2019140 readings
All 88 readings tagged On-balance volume
Also on On-balance volume5 readings
1988Constructing volume-confirmation overlays on OHLC spreadsheet charts1988Cumulative volume trend and 200-session polarity1988Signing volume by open interest change1988Three clocks when cumulative volume led price1989Constructing a volume percentage ratio from advance-decline flow