In this paper, we study Hornets extended with firing probabilities. Hornets are a Nets-within-Nets formalism, that is, a Petri net formalism where the tokens are Petri nets again. Each of these net-tokens has its own firing rate that is independent from the rates of other net-tokens. Hornets provide algebraic operations to modify net-tokens during the firing. For our probabilistic extension these operators could also modify the net-token’s firing rate individually. We use our model to analyse self-modifying systems quantitatively. Hornets are very well suited to model self-adaptive systems performing a MAPE-like loop (monitor-analyse-plan-execute). Here, the system net describes the feedback loop, and the net-tokens describe the adapted model elements. We introduce a sub-class of Hornets that can be translated into Algebraic Nets. Therefore, we can exploit more tools to generate state spaces with probabilities, i.e., in our stochastic setting: discrete Markov chains.