Skip to main content
Track Historical analog comparison
18 / 19
Search the archiveLibrary
Library

2012issue C0261-71

Build a weekly analog as a dated forecast object

A weekly Historical analog comparison becomes a dated forecast after the delay is recorded and the overlay is locked at the correlation-function maximum. Keep an eight-period MACD on that same analog clock so a turning-point reading can be refreshed instead of redrawn by eye.

  • Treat only the broad later path as informative, and let shorter events arrive earlier or later without rewriting the template.
  • Record timing mismatch as a delay in weeks rather than as a change to the forecasting formula.
  • Lock projected and observed weekly prices at the maximum of their correlation function, not by matching short-term swings.
  • Apply the same analog construction to an eight-period MACD panel, and treat lag annotations as part of the forecast object.
Entries in this reading3 entries
Current trackHistorical analog comparison19 readingsLinear regression112 readingsMACD115 readings

What the construction records

Historical analog comparison on the weekly series records delay, locks the overlay of projected and observed prices at a correlation peak, and applies the same construction to a companion eight-period MACD. Chart annotations keep that analog clock on the weekly series as part of the constructed output.

Treat only the broad later path as the template

The analog construction treats only the broad later path as informative. Shorter events may arrive earlier or later without rewriting the template.

Record mismatch as a delay in weeks

Timing mismatch is recorded as a delay in weeks rather than as a change to the forecasting formula. One dated review cites 22 weeks. An earlier 2008 episode is given as 35 weeks.

Identify the present date on the analog

Before the analog can be read as a current scenario, the forecast point that corresponds to the present date must be identified.

Lock the overlay at the correlation peak

The analog is described as reproducing fine structure poorly. The overlay of projected and observed weekly prices is therefore locked at the maximum of their correlation function rather than by matching short-term swings.

DIA weekly analog locked at the November 2011 correlation peak

Weekly DIA actual prices sit on the analog forecast after the overlay is locked at the correlation-function maximum dated 10 November 2011 (the vertical as-of line). A trader should treat only the broad later path as the template: once that lock is refreshed, the next major low is in mid-autumn 2012 rather than August. Values were read from the source plot using its 60–155 dollar scale and the 2005–2016 weekly axis.
Weekly DIA actual prices sit on the analog forecast after the overlay is locked at the correlation-function maximum dated 10 November 2011 (the vertical as-of line). A trader should treat only the broad later path as the template: once that lock is refreshed, the next major low is in mid-autumn 2012 rather than August. Values were read from the source plot using its 60–155 dollar scale and the 2005–2016 weekly axis.DIA · weekly · 2005-01-01T00:00:00.000Z to 2016-10-31T00:00:00.000Z

Position is taken from the correlation-function maximum, not by matching swings by eye. Timing mismatch of 22 weeks as of April 2011, and 35 weeks in 2008, is recorded as delay rather than analog failure.

Refreshing the analog can move a turning point

Separate alignments dated April 2011 and November 2011 each place the projected weekly industrial-average path against the realized path at that correlation-function maximum.

After the November 2011 realignment, the projected price minimum is timed to mid-autumn rather than August 2012. Refreshing the analog can move a dated turning-point reading.

Keep MACD on the analog clock

The same analog construction is applied to a companion indicator panel that includes an eight-period MACD, so the forecast object is not price-only.

Chart annotations on the weekly series record both lag, including a 22-week delay and a 6-week delay, and an 18-week outrunning interval. The analog clock is part of the constructed output.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
18 of 19 in the Historical analog comparison track
201544-48 pp.Next on Historical analog comparisonFrom a drawn price shape to an event-cloud case studyA user can draw or adjust a price shape with connected points, name it, and store it in an event-library for later analog search.
All readings on this track · 19 readings
  1. 1988Crash fear fails the depression regime test
  2. 1990October 1987 cycle overlay and the loss-trap
  3. 1990Constructing nested four-year market cycles
  4. 1991Evaluating quarterly return runs with historical analogs
  5. 1992Evaluating split events across correction and bear regimes
  6. 1993Mining-bullion relative strength as a gold-sleeve regime
  7. 1994A two-horizon case study of a market-breadth oscillator
  8. 1994Extreme short-rate declines as equity regime context
  9. 1997Clustered true-range days as a regime label rather than a top forecast
  10. 2001Nearest-neighbor one-week forecast from log-price patterns
  11. 2001Constructing nearest-neighbor forecasts gated by a trend filter
  12. 2003Regime context for debt-era bear rallies
  13. 2004Testing a 1987 stock and gold analog by wave degree
  14. 2004Shifting calendar regimes and election-cycle analogs
  15. 2006Aligning sugar boom phases with seasonal analogs
  16. 2009Crowd consensus and failed targets as regime context
  17. 2011Treat a long-horizon chart analog as a regime scenario
  18. 2012Build a weekly analog as a dated forecast object
  19. 2015From a drawn price shape to an event-cloud case study
By fieldTime series and forecasting51 readingsBy year201281 readings
All 19 readings tagged Historical analog comparison
Also on Historical analog comparison5 readings
1988Crash fear fails the depression regime test1990Constructing nested four-year market cycles1990October 1987 cycle overlay and the loss-trap1991Evaluating quarterly return runs with historical analogs1992Evaluating split events across correction and bear regimes