Dependence of f_blue on galaxy stellar mass M and cluster-centric distance r/r₂₀₀. Here the model includes only a dependence on M and r/r₂₀₀ (i.e., as Eq. (2), without the Q-related term). Left panel: thick circles with error bars represent the stacked data and the 68% confidence interval. The colour-coded surface represents the fitted model (mean value of the posterior distribution at each [x, y] locus). The model value of f_blue at the [x, y] locus of each stacked data is symbolised by a cross. Right panel: data points with error bars represent the stacked data (slightly shifted along the x-axis for clarity). Solid lines and shaded areas represent the posterior mean and error of the model of Eq. (2). The fit is performed on the 204 individual f_blue measurements.