The path in pattern research

In the initial approach, the optimization process was primarily structural and analytical. The algorithm focused on the geometric relationships within the pattern’s configuration and the statistical accuracy of signal detection. The main goal was to find parameters for which the match between the formal structure of the figure and the actual signal confirmation was the most precise. At this stage, Optuna served as a tool for finding optimal formal characteristics that best matched market conditions. At the next stage, the methodology was expanded — an economic dimension was added to the optimization process. Now each attempt was evaluated not only by formal correctness but also by the model profitability of trades generated from the identified signals. Trading costs, result amplitude, and risk level were considered. Thus, Optuna evolved from statistical parameter fitting to assessing the economic feasibility of a pattern, aligning the process with realistic market conditions. Moreover, while the first version had a single metric — maximizing the share of confirmed signals — the second used a comprehensive evaluation system accounting for profitability, stability, and risk. Optimization became multidimensional, seeking a balance between accuracy, reliability, and growth potential. The focus shifted from finding the best “geometry” to finding the most effective behavioral market response, increasing the model’s practical relevance. In the third stage, the optimization approach became fully risk-managed. While the previous version evaluated signal efficiency only by profitability, the new objective function included risk and stability metrics such as the Sharpe ratio, maximum drawdown, and standard deviation of returns. The system thus considered not only growth potential but also quality of profit, i.e., the relationship between return and volatility. In this approach, Optuna does not simply tune parameters to increase profitability but seeks a balance between profit and robustness of trading results. Optimization became risk-efficiency-oriented, bringing the process closer to the principles of modern portfolio theory. Each experiment is now evaluated as a complete trading simulation, accounting for commissions, stop-losses, and take-profits. This allows not only testing the pattern structure but also modeling its economic behavior under real market conditions. Thus, Optuna becomes a tool that not only calibrates shape but teaches the system to respond optimally to market dynamics. In conclusion, the third version integrates structural, economic, and risk-oriented approaches into a unified system that seeks not the most aesthetic but the most efficient and stable pattern of price behavior. The same optimization framework was applied to data with different time resolutions — 15-minute and hourly. This made it possible to test the stability of the methodology across various observation scales and ensure parameter consistency across different levels of market data aggregation.