One-point matter PDFs beyond TopHat filters
π arXiv:2608.20457 Β· π₯ PDF Β· 2026-08-20 Β· hep-th
Authors: Anton Chudaykin [arXiv Β· scholar] , Alexander M. Kayssi [arXiv Β· scholar] , Sergey Sibiryakov [arXiv Β· scholar]
π Abstract
We study the one-point probability distribution function (PDF) for matter densities averaged with an arbitrary spherically symmetric window function. The PDF is analytically modeled within the path integral framework, enabling a non-perturbative description of large-scale structure. It contains a leading order contribution controlled by the spherically symmetric gravitational collapse dynamics, as well as an order-one factor arising from aspherical fluctuations. We develop a numerical pipeline to compute the leading spherical-collapse part of the PDF and apply it to a family of window functions interpolating between the TopHat and Gaussian filters in coordinate space, as well as to a window function with non-monotonic radial dependence. We find that the PDF weakly depends on the choice of the filter, provided the width of the filter is normalized to yield a fixed linear averaged density variance. For each filter and each value of the averaged density, our pipeline gives the most probable density profile. We find that these profiles vastly differ for different filters in the case of overdensities, but closely follow a universal curve at underdensities. We obtain a perturbative expression for the aspherical part of the PDF valid at small density contrasts. We find from it that all PDFs are equally sensitive to the effective field theory (EFT) corrections accounting for short-scale clustering, regardless of how smooth the filter's boundary is. We test our PDF model against the results of high-resolution N-body simulations. The agreement is excellent for filters with widths larger than 10 Mpc/h. Small discrepancies at a few percent level arise for narrower filters and are interpreted as higher-order perturbative corrections.