Skid resistance markedly depends upon road surface texture. That texture is composed of a range of scales each of which contributes differently to the generation of friction at the tire-road interface. This work aims to contribute to understanding the role of these different scales. The method adopted deploys a signal processing technique, termed Empirical Mode Decomposition, to decompose the texture into a set of component profiles of different wavelengths. The Dynamic Friction Model, a computational friction model already validated on real road surfaces, is then used to determine the relative effect of partially recomposed profiles with their components on skid resistance. The results demonstrate the importance of not only ‘small-scale' and ‘large-scale' textures but also their spatial arrangement and shape. Indeed, on wet road surfaces, ‘small-scale-texture' was found to be key to achieving good skid resistance at low speeds, whilst ‘large-scale-texture' was found to be crucial to maintaining it with increasing speed. But furthermore, the distribution of the summits of the large-scale-textures was established as being able to compensate for a lack of small-scale-texture. Conversely, the reverse was established as also being true, with the small sharp local summits of small-scale-texture being found to compensate for a lack of large-scale-texture.