25-trial experiments on the same data sets as described in Chapter 3 were run on the storing-all-rules version of BruteDL. In each trial, BruteDL with the option set to storing-all-rules was run on the same variously sized training sets and the fixed testing data that are used in the corresponding trial of the experiment described in Chapter 2. For example, we randomly split the iris data set into two subsets -- 30% for testing and 70% for the pre-sampled training data. Then the pre-sampled training data is split into 7 subsets. The first trial of the experiment on BruteDL was run on the first subset of the pre-sampled training data. Before starting the second trial of the experiment on BruteDL, we ran BruteDL with storing-all-rules on the first subset. By running the experiments in this way, the comparison of the results on the original BruteDL against the results on the BruteDL with storing-all-rules will be a good representation of the comparison of the performances of the two versions. The following section shows the results and discussions of the experiment.