Current research on wildlife security games has minimal focus on performance evaluation. The performance of the rangers is evaluated by assessing their game utility, sometimes in comparison with their maximin utility, and other times in comparison with their real-world utility when the game is implemented in a wildlife park. Currently no evaluation framework exists, and this paper proposes an evaluation suite to address this. The movements of the wildlife, the rangers, and the poachers are simulated over a grid of cells corresponding to the wildlife park, where cells containing geographical obstacles are excluded. Poaching and arrest frequency are the primary evaluation measures used. Firstly, we develop a null game to act as a baseline. Typically, one would expect random behaviour of all agents in the null game. However, we simulate random movement for the rangers but more intelligent movement for the poachers. The motivation for this design is to assess whether executing the Stackelberg game yields significantly better ranger performance than random movement, while keeping the poachers’ behaviour consistent. The intelligent poachers move by taking their geographical preferences into account and learn from poaching and arrest events. Secondly, we propose that the rangers act as the Stackelberg follower instead of the leader. We formulate a simple pure-strategy Stackelberg game and implement four variations of the game within the framework. The results of the simulations show that the rangers perform better than random when using the Stackelberg game and perform best when acting as the follower.