In the project that is the basis for this report we have investigated how successfully a computer that uses Q-learning can learn to play the game of Alquerque. The computer has been trained against a greedy and a randomized player. Different parameter settings for the Q-learning agent have been tested plus some modifications as the implementation of an eligibility trace. Some settings have proven truly successful in beating the greedy AI but all tests against the randomized player have shown to be inconclusive.