The concentration-mass relation is dark matter structure's growth chart: in ΛCDM, halos that form earlier are denser because the universe was denser, so lower-mass halos should be more concentrated, with a tight relation and modest scatter emerging from gravity-only physics (Bullock and Boylan-Kolchin 2017).
The measured chart is scattered beyond recognition. Dwarf galaxies show low-concentration cores where the relation demands dense cusps; some massive clusters run startlingly over-concentrated, with Einstein radii and lensing efficiencies their assembly histories should not produce; and at fixed mass the spread of measured concentrations dwarfs the simulated scatter (Kaplinghat et al. 2016). The diversity problem generalizes the core-cusp problem: it is not one systematic offset but a variance the gravity-only physics cannot generate, since mass and accretion history are the only dials the model owns. The repairs each add a sector: baryonic feedback tuned per system, self-interacting dark matter with cross-sections chosen to thermalize cores while sparing clusters, or warm dark matter trading small-scale structure away.
The standing is a variance problem, which makes it sharper than an offset: whatever sets halo structure has more degrees of freedom than the standard model's two, and every proposed third dial belongs to a different theory.