freedom to tune light transport properties in these media. Indeed, shapes such as prolate ellipsoids or cylinders can be packed up to higher densities delaying the onset of spatial correlations at the cost of increased angular correlations. [10][11][12] Interestingly, both these aspects-namely, the high density and prevalent orientation exhibited by packed rods-can contribute to increase the overall turbidity, which makes cylinders particularly suited to realize highly turbid materials with a flat response over a broadband wavelength range. [13] Indeed, the smallest possible transport mean free path for a given refractive index contrast reported so far in the visible range has been obtained with GaP nanocylinders [14] and, for lower refractive index materials, claims of optimized scattering exist for the chitinous network structure of the Cyphochilus beetles. [5,13] In recent years, the latter has inspired an array of bio-mimicking materials attempting to reproduce its outstanding efficiency in terms of strong light scattering and limited material usage, taking advantage of a wide range of fabrication techniques including electro-spinning, super critical CO 2 foaming, polymer phase separation, and direct laser writing, [6,7,9,[15][16][17] to name a few.However, as opposed to nanoparticle systems, network materials are characterized by several additional aspects other than number density and spatial correlations, which makes it difficult to understand what key parameters should be optimized to design highly scattering network structures.Few notable examples include phase percolation, angular correlation, and network valence, all of which concur in determining their scattering properties. [13,[18][19][20] In this respect, simple generative models for photonic structures allowing to investigate the effects of these parameters separately are much needed to gain insight on their role and relevance.In this work, we describe a simple branching random walk (BRW) algorithm to generate random network structures inspired by that of the Cyphochilus beetle. Notably, the model allows to control and vary independently the volume fraction and degree of angular correlations without altering structural parameters such as the slab thickness, the shape, and aspect ratio of the constituent elements. The optical and transport properties of the generated structures have been investigated through finite difference time domain (FDTD) calculations and an inverse Monte Carlo (MC) approach, showing that the bright reflectance of the Cyphochilus beetle can be easily matched and surpassed by acting solely on the degree of anisotropy and the volume fraction of the network. To the best of our knowledge, this is the first rigorous demonstration of the key role played by structural anisotropy in highly reflective disordered samples.3D disordered networks are receiving increasing attention as they represent a versatile architecture for highly scattering materials. However, due to their complex morphology, little is known about the interplay betwee...