Strong lensing arcs are gravity drawing its own portrait: background galaxies stretched into giant arcs, Einstein rings, and multiple images by the mass of foreground clusters, with every arc's position, radius, and shape a direct constraint on the lensing mass distribution. The portraits keep disagreeing with the sitter. Giant arcs with length-to-width ratios beyond 10:1 are more abundant than ΛCDM simulations produce, the arc-statistics problem that has resisted resolution since the 1990s (Bartelmann et al. 1998); arc radii and positions conflict with mass models built from weak lensing and X-ray data for the same clusters; and the detailed morphologies, radial arcs, tangential arcs at unexpected radii, anomalous flux ratios in multiply-imaged systems, demand mass distributions more complex, clumpy, and extended than smooth NFW halos provide (Meneghetti et al. 2013).
The standard repairs trade one problem for another: raising halo concentrations to make more arcs conflicts with the concentration-mass relation; adding substructure helps morphologies but the required clumpiness exceeds simulated subhalo populations, the same direction as the galaxy-galaxy strong lensing excess; and line-of-sight projection arguments require chance alignments at statistically uncomfortable rates. The lensing sky is simply more efficient and more textured than the simulated one.
The standing spans three decades of arc surveys, sharpening as JWST and Euclid resolve arc systems in unprecedented detail across hundreds of clusters.