First column: ‘Raw’ O−C diagrams, namely, without subtracting the quadratic fit performed by Benkő et al. (2019). Note: there are very different vertical scales shown in each panel. Second column: Pseudo-residuals (see formula (9)) associated with the best statistical model. Third column: Normalised distribution of pseudo-residual values (blue histograms) compared to the standard normal distribution (orange curves). In the upper right corner of the panels, the resulting optimal model (M1-M4) is indicated. In the upper left corner, the p-value of the normality test is given.