Lower performance using parameters found by cross validation than default parameters

Question

I divided the dataset into 2/3 for training set and 1/3 for test set. Then I apply cross validation for the selection of the parameters on the training set. Then I train on