We present a new method to asses the strength of indirect interactions and to indentify candidate keystone species in quantitative food webs. We apply this method to the structural analysis of a host-parasitoid community. The strength and symmetry of indirect interactions between 12 leaf-miner hosts and their 27 hymenopteran parasitoids are quantified. It is shown that (1) quantifying longer pathways helps in determining which species have more important direct or indirect effects on others, (2) a keystone pattern of relative species importance, based on positionality in the interaction network, seems to characterize this community, (3) considering longer pathways results in a characteristic “few strong - many weak” distribution of interaction strength, and (4) between the majority of species pairs the interaction is weakly asymmetrical. We emphasise that a very simple network algebra approach may offer important predictions on both species- and community-level patterns.