The challenge is how to identify how many one-hop neighbors that node n x could help via itsforwarding. Since each node should generate a recommendation for each lost BSM that it has explored, the number of generated recommendations can be used for inferring.