Brute-Force Pattern Search with Polynomial Signals and Smoothness Filters
Summary
This article extends a brute-force method for finding trading patterns in currency data. It expands a polynomial signal’s inputs from close-open changes to several combinations of each bar’s open, high, low, and close, allowing coefficient combinations to express additional price relationships. A linearity measure is added to rank candidate equity curves by smoothness alongside profit factor or expected payoff. This filter requires another calculation pass and slows the search, while the richer polynomial also increases computational cost.
The author frames a trading signal as one continuous-valued function rather than a set of separately tuned Boolean conditions. Stronger signal thresholds may reduce trade frequency while selecting observations believed to be more predictable. The article gives a conceptual efficiency model linking the expected number of qualifying strategies to search frequency, sample duration, minimum trades, and linearity constraints. It offers only limited empirical coverage: a few currency pairs and one timeframe, constrained by computing resources. The author cautions that small samples can produce chance results and that finding a profitable historical pattern does not establish its future durability.
Key ideas
- The method searches polynomial combinations of bar price differences to generate continuous-valued trading signals.
- A linearity filter ranks smoother equity curves but adds computational cost.
- Increasing the required signal strength can reduce trade frequency while selecting more predictable observations.
- Search efficiency depends on sample duration, quote data, trade-count requirements, and any smoothness constraint.
- Small samples and historical profitability do not establish that a pattern will persist.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.