Disordered carbons offer promise of performance gains relative to graphite as the anode host material in Li ion batteries. They also present major challenges in determining structure-property relationships, in particular the location, density and properties of the site (or sites) at which reversible Li uptake occurs. Our understanding of graphite anodes is built on 35 years of research on the crystal compound LiC6. Here we review our efforts to obtain similar understanding of disordered carbons, using local structure determination via radial distribution function analysis, quantum chemical simulations and neutron vibrational spectroscopy. The ultimate goals are to understand the origin of the very high capacity for Li uptake exhibited by some of these materials, and to establish a scientific basis for optimizing their performance in real batteries. Two illustrative examples will be discussed in detail: the effect of residual hydrogen and ＇edge sites＇ on Li binding, and the detailed evolution from long-range to short-range order in ball-milled graphite.