Research addressing the transmission of violent offending has predominantly considered the spatial and temporal scales at which crime risk is transmitted. However, geographical features are only a part of the explanation of how crime is transmitted. Social processes may also impact the transmission of crime. In response, the role of person-to-person interactions within offender networks and adopting a social network approach is increasingly argued to be key to understand how crime risk is transmitted among individuals. However, extant research focuses on victimization within co-offending networks. The transmission of violent offending between offenders has been considered to a lesser extent. Therefore, I study how physical violent offending is transmitted between individuals within co-offending networks. For this, I adopt a network-based methodology that allows me to track the person-to-person transmission of physical violent offending in a social network of offending peers. Using police recorded crime data, the structure of the co-offending network of physical violent offenders is mapped, which enables me to establish the level of connectivity between the offenders. Subsequently, I apply network centrality measures to determine to what extent offenders in a network are important in respect to their network position. Finally, the predictive power of social network measures is compared to a demographics model, to determine the probability of future physical violent offending.