The aim of this study is to define whether using exercise-based games increase the performance of learning. For this reason, two basic questions were tried to be answered in the study. First, is there any difference in learning between the group that was given exercise-based games and the group that was not? Second, is there any difference in learning between the group that used exercise-based games at end of the process of learning and the group that was not applied this but taken the questions of exercises in game material? This research has been conducted within the subject of Testing and Evaluation in the program of Kocaeli University Primary Maths Teacher's College. Experimental design with a pre test-post test control group was used in this study. Experimental process based on game material was used in 120 minutes at the end of a 3-week-teaching period. The reliability values (KR-20) of the two tests were found to be .79 and .71 which were used to evaluate learning level. The study has reached a conclusion that game materials used at the end of learning process have increased the learning levels of teacher candidates. However, the similar learning levels have been observed among students who were taken printed exercises instead of using learning game method to reinforce the traditional learning in the research. This means that in method of applying teaching games in addition to the traditional teaching, there isn't any difference of learning efficiency of students answered the questions based on competition and fun and the group who only answered the questions. This study is expected to contribute defining in which situations games are effective.