I am using brier score as my scoring metric, as opposed to log loss. my reason for doing so was because i read it was more focused on the positive class than log loss. 
But how is it? 


https://machinelearningmastery.com/tour-of-evaluation-metrics-for-imbalanced-classification/