1: \begin{definition}[Incomplete information  sampled network game]

2:   An incomplete information  sampled network game $\mathcal{G}^{in}(\mathtt{S},U,W)$, is a  network game  with a random number $N$ of agents, whose  types $\{t^i\}_{i=1}^N$ are sampled independently and uniformly at random from $[0,1]$, that  interact according to a network $P^{[N]}_s$ sampled from the graphon $W$ according to Definition \ref{sample}.\footnote{The distribution of  $N$ does not matter for our results, with the exception of Theorem \ref{thm:bayes} where we assume that the support of such distribution is bounded from below by $N_\textup{min}$.} Each agent $i$ has information about the graphon $W$, his own type $t^i$,  the strategy sets $\mathtt{S}$ and the payoff function $U$, while is uninformed about $P^{[N]}_s$ and the other agents types $t^{-i}$.

3: \end{definition}

4: