Skip to main content
Track Linear regression
42 / 43
Library

2019issue C0362-68

Pair-trade layer construction versus average-spread management

Pair construction can add discrete layers at successive spread prices and treat the average-spread-price only as a retrace checkpoint for production, not as the working position.

  • Pair construction can add discrete layers at successive spread prices rather than treat the blended average-spread-price as the working position.
  • In the worked example, layers at -1, -2, and -4 give an average-spread-price of -2.33 that is used only as a retrace checkpoint.
  • Closing the -4 layer at that average would have been production and freed capital; sitting through the round trip left the book fully loaded, with risk described as having increased.
  • A noise-pair is the construction case for scaling. A signal-pair is the case for a one-off-trade. Check a layer against catalyst, average-daily-range or average-true-range, and other readouts rather than the latest bounce.
Entries in this reading3 entries

Layers instead of a blended working position

Pair construction can proceed by adding discrete layers at successive spread prices rather than by treating the blended average as the working position. A layer is a discrete pair-trade unit opened at its own spread-price, the quoted difference between the two legs at that moment, and managed on its own merits.

The average-spread-price is the capital-weighted mean of all open layers. It is useful as a checkpoint. It is not the trade trigger and it is not the position that is being worked.

The average-spread-price as a production checkpoint

In the worked example, three layers at spread prices -1, -2, and -4 produce an average-spread-price of -2.33. That average is then used only as a retrace checkpoint.

Acting at that average by closing the -4 layer would have realized 1.67 times that layer's size. That close is production: a realized result on one layer while other layers may remain open. Capital freed by the close could later be redeployed if the spread returned to -4.

Sitting through a move from -4 back to the average-spread-price and then back to -4 leaves the book fully loaded. No production is closed. Risk is described as having increased.

Context for a layer decision

A layer decision is supposed to be checked against catalyst, average-daily-range or average-true-range, and other indicator readouts rather than against the most recent bounce alone. Average-daily-range is a lookback measure of typical pair movement used to judge whether a retrace is large enough to act. Average-true-range is a related volatility lookback used with other indicators to keep the decision in context.

Noise-pairs, signal-pairs, and spacing

Pairs classified as range-bound, the noise-pair case, are the construction case for scaling. Pairs classified as trendy, the signal-pair case, are the case for one-off entries. A one-off-trade is a single-layer entry that is not scaled.

When adding layers, a logarithmic spacing of entries is presented as potentially preferable to fixed-distance scaling, with size allowed to change at inflection points.

Starting size, stops, and more names

Initial construction, smaller size per idea and, in pairs, reduced direct market exposure, is offered as a way to lessen the need to stop out on ordinary fluctuations.

A stop is framed as a stop-as-validity-check for staying in, adding to, or exiting a pair, including an averaged-in book, rather than only as a device that cuts a loss. It tests whether the original thesis still holds.

Spreading risk across more qualified symbols, illustrated as ten names instead of one, is presented as reducing the need to be exactly right on a single construction.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
42 of 43 in the Linear regression track
202050-56 pp.Next on Linear regressionA convolution slope built from nested linear regressionThe projection step is a linear-regression value of the chosen price series over a user length, evaluated one bar ahead.
All readings on this track · 43 readings
  1. 1990Constructing dollar baselines from rates, inflation, and residuals
  2. 1990Constructing a nominal index value from forward earnings and fitted yield
  3. 1990Constructing a nominal index price from earnings and a fitted yield
  4. 1990Endpoint-pinned price paths are not forecasts
  5. 1990Constructing least-squares polynomial smoothers
  6. 1991Endpoint growth rates versus linear-regression consistency
  7. 1991Out-of-sample checks for linear growth fits
  8. 1991Trend as persistence, not a straight line
  9. 1991Quadratic trend, residual oscillator, and a secondary cycle calendar
  10. 1991Time-origin offset and residual-price divergence on a quadratic least-squares fit
  11. 1991A least-squares trendline from ordered prices
  12. 1992Constructing log-linear growth and reliability screens
  13. 1992Constructing log-linear growth-rate baselines
  14. 1992Next-session high, low, and close from rolling linear regression
  15. 1993Auditing an index price-earnings multiple with short-rate regression
  16. 1994Regression-seeded nested exponential price filter
  17. 1994Constructing the double exponential average from lag cancellation
  18. 1994Evaluating money supply as a linear leading-index baseline
  19. 1995Constructing least-squares trend channels
  20. 1995Linear baseline holdout checks for annual bill-rate forecasts
  21. 1995Projection bands from high and low regression slopes
  22. 1995Evaluating a least-squares end-point moving average on a known test series
  23. 1996Constructing an endpoint moving average from a least-squares line
  24. 1996Scoring equity path consistency with a k-ratio overlay
  25. 1996Evaluating month-end yield gaps for equity regimes
  26. 1996Constructing session-indexed standard error bands
  27. 1998Evaluating linear regression baselines for index valuation
  28. 1998R-squared as a two-state trend filter from a price-time fit
  29. 2000Second-order moving-average lag correction
  30. 2002Price regression line versus beta for index tracking
  31. 2003Regression slope with an r-squared trend confidence gate
  32. 2003Constructing finite-volume-element divergence with slope comparison
  33. 2004Building a daily score from regression, retracement, and volume
  34. 2004Constructing least-squares trendlines from ordered prices
  35. 2007Rectangle breakout targets beyond height
  36. 2007Confirming a price trend with regression slope and r-squared
  37. 2008A linear-regression angle assembled as one trend filter
  38. 2010A two-state swing machine from four running extremes
  39. 2016Score oil-complex tightness before divergence or regression
  40. 2017Nikkei-yen intermarket divergence as a regime case study
  41. 2017Constructing Calmar ratio and linear regression baselines
  42. 2019Pair-trade layer construction versus average-spread management
  43. 2020A convolution slope built from nested linear regression
All 112 readings tagged Linear regression
Also on Linear regression5 readings