Non-competent examples (as stated in [46]): In OVO strategy, all classifiers are not trained with all the instances in the data set, but in testing phase, the new instance is submitted to all classifiers. The classifiers that have not been trained with the instance from the class of the new example will make a prediction that probably affects negatively the final results since these classifiers are not competent. Hence, detecting which classifiers are giving a response without really knowing anything about the example that has been introduced should improve the behaviour of OVO scheme.