Duelist Algorithm for Population-Based Strategy Optimization
Summary
The article presents the Duelist Algorithm as a population-based method for optimizing strategy parameters. Candidates compete in randomized pairwise duels; the better-performing candidate usually wins, while a luck term can affect the outcome. The algorithm then gives losers a chance to copy coordinates from winners, allows winners to mutate selected coordinates, and creates additional candidates near champion solutions. After these steps, weaker candidates are removed to restore the population size.
The discussion describes the algorithm’s parameters, initialization, iteration flow, and MQL5 implementation, then reports comparisons on benchmark test functions. The stated strengths are speed and relatively low variation across tests, while the stated limitation is weaker convergence accuracy. The article does not show that optimizing synthetic functions translates into profitable trading or robust out-of-sample strategy settings. Its results depend on the benchmark suite and implementation choices, which the author notes may differ from canonical versions of optimization algorithms.
Key ideas
- Losers learn by adopting selected parameter values from winners, while winners can explore through random changes.
- Champions mentor new candidates near their parameter settings and are excluded from duels.
- A luck term makes duel outcomes probabilistic, and survivor selection keeps the population bounded.
- The reported benchmark tests characterize an optimizer, not the trading performance of strategies it might tune.
- The article identifies speed and low test-result variance as strengths, with lower convergence accuracy as a limitation.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.