k-nearest neighbour cumulative distribution functions (kNN-CDF, left panel), count-in-cylinder (CIC, middle panel), and void probability function (VPF, right panel) of the FLAMINGO simulation (symbols) and the SHAMe model (lines). The results for BOSS-like, shown in red, are displaced along the y axis to facilitate the comparison. The different line styles represent the statistics used to fit FLAMINGO as labelled. The lines represent the median of 1000 MCMC points when fitting the projected correlation function and the galaxy-galaxy lensing (dotted lines), as well as the projected correlation function, multipoles of the correlation function, and galaxy-galaxy lensing (solid lines). It is worth noting that none of the statistics shown in this figure were used in the empirical model fitting process. The shaded region corresponds to the 16th and 84th percentile of the distribution (only shown for the solid line). Only the predictions for N_k = 2 and 5 are shown for the kNN-CDF. Similarly, the CIC only shows the predictions for r_CIC = 1 and 10 h⁻¹ Mpc. The bottom panels show the ratio between the empirical models and FLAMINGO.