Genetic Optimization of Weighted Votes in a Three-Strategy Ensemble

Summary

This article adds a Williams Percent Range reversal strategy to an Expert Advisor ensemble that also contains RSI momentum and moving-average crossover strategies. It describes a weighted voting policy: each strategy contributes a vote with a tunable weight between zero and one, and the genetic optimizer adjusts those weights alongside other strategy parameters. A weight near zero can indicate that a strategy contributes little in a particular configuration, while other configurations may benefit from all three.

The article frames this as a way to explore whether every strategy is useful, rather than assuming the ensemble needs all its components. It notes that vote weights are not constrained to sum to one and that the answer depends on the selected settings. The supplied text omits detailed optimization results, and the author’s earlier class-validation results are also omitted. The findings therefore offer a configuration-dependent selection approach, not general evidence that the ensemble or any strategy will remain profitable.

Key ideas

• The ensemble combines RSI momentum, moving-average crossover, and Williams Percent Range reversal strategies.
• Each strategy receives an independently adjustable vote weight between zero and one.
• A genetic optimizer tunes vote weights to explore whether a strategy helps under a given configuration.
• The weights are not required to sum to one, so they represent relative contributions rather than normalized portfolio shares.
• Whether all strategies are useful can vary with the chosen settings, and the supplied text omits detailed results.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.