More than 1 track per particle and tracks far from cluster

[doloresgarcia]

• two tracks that belong to the same particle, atm we are not really dealing with this because we set all tracks to have a high beta and become condensation points..
• tracks that should not be associated to the cluster because they are bad

Both cases could be labelled during training:

• the two tracks case by selecting the one with the lowest chi square or closest to the cluster (ECAL)
• the second one by cluster to the cluster
  but then we can't have tracks be condensation points

not sure how atm the clustering is dealing with this double tracks

another way would be to cut after training but then we need to check what the training is doing now