Research on civil war termination typically classifies conflict outcomes into homogeneous categories. Civil wars are conventionally described as ending in ‘victory’ for one side, ‘negotiated settlement’, or ‘stalemate’. However, each of these common categories conceals a significant degree of unexplored variation. Not all peace processes are the same, nor should we assume equivalence between all ‘victory’ outcomes. Ignoring the variability in outcomes obscures our understanding of how civil wars actually end and can undermine efforts to examine how this may influence the sustainability of the ensuing peace. This article focuses on the concept of rebel victory and argues that it has been used to describe a wide range of civil war outcomes that share some features but that also differ along three dimensions: (i) the residual threat posed by the defeated regime; (ii) the role of external actors in enabling victory; and (iii) rebel fragmentation. The article describes these dimensions and the influence that they can exert on the likelihood of continued political instability in the aftermath of rebel victories. The examples of rebel victories in the Central African Republic (2003 and 2012) are used to demonstrate the analytical utility of this novel conceptualisation of rebel victory.