In this paper, we address the challenge of believability in multiplayer video games. Our contribution is a system for assessing the believability of computer players. The state of the art examines existing methods and identifies seven distinguishing features that differ considerably from one assessment to the next. Our investigation reveals that assessment procedures typically alter gameplay, posing a considerable danger of bias. This is a major flaw since computer players are evaluated in a specific context rather than in the context of the game as it should be played, potentially skewing the findings of the evaluation. As a result, we begin on a trial-and-error process, with each new proposal building on the achievements of the previous one while removing the flaws. New proposals are tested with new assessments, a total of three experiments are then presented. We created a computer program that partially automates the execution of the assessment procedure, making these trials easier to implement. At the end, thanks to our proposal, gamers can assess the believability of computer players indirectly by employing reporting forms that alert users to the presence of bots. We assume that the more a bot is reported, the less credible it becomes. We ran a final experiment to test our proposal, which yielded extremely encouraging results.