“…A good way to observe the emergence of complexity is to perform laboratory earthquakes in controlled conditions, starting from a homogeneous initial state, and to monitor the spontaneous development of seismic cycles (Aubry et al., 2020; Bayart et al., 2018; Bolton et al., 2020; Dresen et al., 2020; Goebel et al., 2013; Guérin‐Marthe et al., 2023; Kandula et al., 2019; Leeman et al., 2016; Li & Zhou, 2021; Marty et al., 2019; McLaskey, 2019; Passelègue et al., 2016; Scuderi et al., 2017; Sobolev et al., 1996; Xu et al., 2019). This can be nicely complemented by numerical simulations, provided that they are able to reproduce the main physics at stake, which include rock elastodynamics, damage mechanics, and granular physics (Casas et al., 2022, 2023; Dorostkar et al., 2017; Guo & Morgan, 2007; Mair et al., 2002; Mollon et al., 2021; Morgan & Boettcher, 1999; Papachristos et al., 2023; Taboada & Renouf, 2023; Wang et al., 2019). In the present paper, we revisit numerical data originating from a set of simulations described in details in Mollon et al.…”