Right now, the selection of targets from the set of triadic_closure_candidates is biased towards those with more connections in common (i.e., weighted by friends-in-commom).
We need an alternative version where nodes are selected uniformly at random by adding a flag in the constructor of netin.TriadicClosure