Skip to main content
Track Contrarian strategy
14 / 20
Search the archiveLibrary
Library

2005issue C041-5

When trend rules need a losing filter

A Mechanical trading system is complete only when entry, exit and abstention are specified as one procedure. This archive note follows the case for Robustness testing, a Contrarian strategy on losing background breakouts, and designer-controlled parameter search after similar trend rules become crowded.

  • Specify entry, exit and abstention together so a Mechanical trading system can be tested as one procedure rather than as separate discretionary judgments.
  • Robustness testing asks a rule only to be passable in an old favorable era, a middling multi-year stretch and a recent bad year, not optimal in any single window.
  • A Contrarian strategy takes a longer breakout only when a shorter background breakout is losing, a mark meant to show consolidation and forced liquidation.
  • Keep parameter ranges under the designer's control and prefer selective, infrequent trading over densely parameterized short-horizon index procedures.
Entries in this reading3 entries
Current trackContrarian strategy55 readingsRobustness testing80 readingsMechanical trading system322 readings

Keep the procedure in one package

A Mechanical trading system is treated as complete only when entry, exit and abstention are specified together. That package lets a designer test the rules as one procedure rather than as separate discretionary judgments.

When similar rules crowd the tape

The interview argues that markets became harder for trend-following procedures because more participants used similar rules. In that account, the market's job is to derail those systems.

A candidate volatility screen

Volatility screening is presented as a candidate refinement among trend-following candidates. The lower-volatility trades are treated as the more usable subset, even if further narrowing is an idea rather than a data result.

Robustness testing across unlike windows

Robustness testing is described as requiring a rule to have been merely passable in an old favorable era, a middling multi-year stretch and a recent bad year, rather than optimal in any single window.

Balance from the other sleeve

Trend-following and countertrend components are paired so that taking one kind of trade also produces the other. The pairing is meant to create a balance that can work when the trend-following sleeve does not.

When a longer breakout is allowed

A Contrarian strategy is specified for n-day breakouts: a longer breakout is taken only when a shorter background breakout is losing. That condition is meant to mark consolidation and forced liquidation rather than an already-trending tape.

Parameter search is kept under the designer's control. Ranges are assigned first, then any promising region is re-run repeatedly to narrow optimization rather than letting programmers invent numbers.

Short-horizon index procedures are criticized as easier to discover and quicker to be arbitraged away after they leave paper. Selectivity and infrequent trading are preferred over densely parameterized short-term systems.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
14 of 20 in the Contrarian strategy track
20061-4 pp.Next on Contrarian strategyEvaluating sector breadth as a contrarian regime filterDefine the daily sentiment-count as how many names in a fixed list of fifteen sector funds close below their own 25-day moving averages, so the reading is always an integer from zero to fifteen.
All readings on this track · 20 readings
  1. 1986A pre-trade checklist for crowded trading methods
  2. 1988Contrary decision process versus lagging bullish consensus
  3. 1988Constructing contrarian signals from polled consensus
  4. 1989Constructing a ten-input breadth diffusion score
  5. 1989Even-number first penetration as a construction problem
  6. 1991Joint strength and sentiment rules for entries and exits
  7. 1995Evaluating contrarian sentiment thresholds for bond futures
  8. 1998A weekly composite-score that classifies despair, euphoria, or neither
  9. 1998Lagged phase clocks for market cycles
  10. 2003Process rules when a correct view fails
  11. 2003Closing stops and equity-curve filters as one decision process
  12. 2003One procedure for entry, exit, and standing aside
  13. 2004Cross-index volatility gauges as a premium regime map
  14. 2005When trend rules need a losing filter
  15. 2006Evaluating sector breadth as a contrarian regime filter
  16. 2011Sideline patience versus crowd noise
  17. 2012Assembling a sentiment zone oscillator as one contrarian filter
  18. 2012Currency regime context for fading rallies
  19. 2014When late trend confirmation is a sideline signal
  20. 2017Write a daily contrarian rule from the closed-end-fund discount
By fieldSystem design203 readingsBy year2005167 readings
All 55 readings tagged Contrarian strategy
Also on Contrarian strategy5 readings
1986A pre-trade checklist for crowded trading methods1988Constructing contrarian signals from polled consensus1988Contrary decision process versus lagging bullish consensus1989Constructing a ten-input breadth diffusion score1989Evaluate a contrary put-call premium ratio at a stated horizon