1988issue C071
Cycle rule inputs until the parameter search stops
A multi-parameter search can score a starting rule-input set, vary each input in turn, and halt when a later cycle leaves every value unchanged. A single-parameter hunt can step while the score improves, then reverse with a shrinking increment until the step is small enough to stop.
- A multi-parameter search can begin with a user-specified rule-input set, score that starting combination on a historical quote file, and vary the first parameter until a preferred value is located.
- After each remaining parameter is varied in turn, a parameter cycle returns to the first input and repeats until a later cycle leaves every parameter unchanged.
- A single-parameter hunt can step a value while the score improves, then reverse and apply a shrinking increment after the score stops improving.
- Search stop is reached when further cycling leaves every parameter unchanged, or when a single-parameter increment becomes small enough to halt.
Begin with a scored rule-input set
The motivating design case is a combination of ten trading parameters evaluated over an eight-year quote file. Those numeric values form one rule-input set: they jointly define entry, exit, and abstention behavior for a single testable procedure.
A multi-parameter search can begin with a user-specified rule-input set. That starting combination is scored on a historical quote file. The first parameter is then varied until a preferred value is located.
Repeat a parameter cycle until none of the inputs change
After the first parameter is set, each remaining parameter can be varied in turn while the others stay fixed. The procedure then returns to the first parameter and repeats the cycle.
The multi-parameter search can be treated as finished when a later cycle leaves every parameter value unchanged. That unchanged pass is a search stop.
Shrink the increment when the score turns down
A single-parameter hunt can score the current value, then step that value by an increment while the score improves. When the score stops improving, the hunt reverses direction and uses a halved increment.
Halving the increment at every direction change lets the single-parameter hunt stop once the step size becomes small enough. That collapsed increment is the other form of search stop.
Count score passes through the quote file
The supplied walkthrough required ten score evaluations to locate one parameter value. Repeated evaluations meant only seven full passes through the quote file.
Those seven score passes settled on an integer parameter value of 27.
All readings on this track · 23 readings
- 1984Evaluating mechanical systems beyond net profit
- 1987Mechanical systems need more than pretty charts
- 1987Evaluating silver moving averages and RSI by scoring rule
- 1988Simple systems over optimized complexity
- 1988Cycle rule inputs until the parameter search stops
- 1990Split-sample tests for stochastic parameter grids
- 1993Four-subsystem trading system construction
- 1993Randomized price paths for stock system parameters
- 1994Mechanical timing with out-of-sample and filter checks
- 1995Evaluating asymmetric percent-reversal trading rules
- 1996Evaluating a daily volume-strength and new-high new-low timing procedure
- 1997Rule-based intraday system design with genetic search
- 1997Evolving three-rule systems with genetic templates
- 1997Audit the holdout and the exposure before the fitted entries
- 2001Inner weights are part of the candlestick signal
- 2003Mechanical short-term rules for news-sensitive rallies
- 2004Band construction as a testable trading procedure
- 2006Layered day-session automation from brackets to rules
- 2006Why mechanical rules must be retuned per security
- 2006A linear-regression next-bar forecast needs a searchable cutoff
- 2015Construct if-then rules, then backtest and optimize
- 2015Mechanical system design as a critique of discretion
- 2025Joint channel design with a trend filter and stop-loss orders