In competitive electricity markets, prices are determined by the collective behaviour of suppliers and consumers. Hence, these systems rely on the balance between supply and demand, and sudden changes in the underlying conditions can lead to significant price fluctuations. In the face of the recent transformations in Italian electricity markets, which are driven by an increasing share of non-programmable renewables, energy crises, and geopolitical tensions, our study focuses on effective forecasting methodologies. We compare kernel and linear regression for predicting market equilibrium prices, both in point and probabilistic sense. We showcase the potential of both linear and non-linear models when carefully engineering the problem of interest, which involves properly selecting data and variables. The noteworthy outcome is that, while linear and non-linear models may differ in nature, their performance converges closely, attesting to the robustness of our approach in achieving reliable forecasts. We describe data sources and assumptions in exploratory univariate analyses, and the performance of the final multivariate model is evaluated over a test period on September 2023.