How renewable energy mix shapes day-ahead price predictability: A comparative analysis of solar, wind, and biomass in Czechia

Understanding how renewable energy mixes influence day-ahead electricity price predictability is essential for market operations and renewable integration in thermally dominated systems. This study analyzes one year of hourly data from the Czech electricity market. Five models are evaluated across four renewable-dominance regimes to assess price predictability under observed system conditions. Linear Regression provides the best overall predictive performance across the full sample, while Random Forest is used for regime-specific analysis due to its ability to capture nonlinear interactions. Solar-dominated hours yield the lowest forecasting errors, with Random Forest achieving an RMSE of 10.97 euros per megawatt hour, reflecting the regular diurnal alignment of solar generation with load. Fossil-dominated hours exhibit the highest errors, with an RMSE of 48.53 euros per megawatt hour and frequent extreme price spikes. In comparison, solar-dominated conditions reduce RMSE by approximately 77%. Biomass contributes moderate stabilizing effects, while wind plays a minimal role in price formation. SHAP analysis applied to the Random Forest model shows that higher solar and overall renewable shares are associated with reduced sensitivity of prices to imports and load, alongside persistent lagged-price effects. Seasonal and weekly error patterns confirm that predictability improves when renewable output aligns with demand. The mechanisms identified are applicable to similar thermally dominated, interconnected power systems beyond Czechia. The findings indicate that expanding solar and biomass capacity can stabilize day-ahead prices in flexible thermal systems and support regime-aware decision-making for market operators.