Forecasting Market Values of Players in Europe’s Top Five Leagues Using ARIMA and EA FIFA Attribute–Augmented Models

Abstract

This study examines how professional football players' market values change over time using monthly data from Transfermarkt and performance attributes from FIFA23–FIFA25. It uses two related time-series approaches. First, univariate ARIMA models trace the market value path of each player separately. Second, ARIMAX models build on that baseline by adding FIFA-based attribute groups as exogenous regressors to explain short-term changes in value. Treating each player as an independent time series makes it possible to capture different career paths instead of reducing them to aggregate patterns. The results suggest that most players' market values move in relatively smooth and stable ways, and low-order specifications such as ARIMA (1,0,0) appear most often. Adding FIFA attributes makes the model easier to interpret and shows positional differences. Attacking players are more responsive to attacking and skill-related attributes, whereas defenders and goalkeepers react more to defensive and mentality-related attributes. A few players do not fit these patterns well. Their valuation dynamics seem to be shaped more strongly by transfers, sudden reputation shifts, or career-stage effects. Taken together, the study offers a quantitative framework for linking time-series modeling with performance attributes to understand football player market values more clearly.