We reproduce our measurements with AbacusSummit n-body simulations and come up with an analytic prediction. The scale-dependence of the signal traces the underlying tidal field and its amplitude is a result of small-scale (<1 Mpc) dynamics.

This compliments traditional 2-point clustering by retaining directional information imprinted by tidal forces, and contains additional LOS information compared to weak lensing. However, any applications will require some sort of calibration/understanding of the small effects!