Computer games are quickly gaining momentum by enabling new approaches to teaching and learning experience for programming courses in K-12 curriculum. However, it remains unclear if the game interface and elements created by using three-dimensional (3D) virtual worlds combined with visual programming languages or a visual programming environment can affect students' learning performance. This quasi-experimental study presents evidence about how a game can assist boys and girls to gain a greater understanding on skills related to CT for developing, implementing and transforming their solution plans into code based on their computational problem-solving strategies. A total of fifty (n = 50) high school students who volunteered to participate in this study divided into a control group (n = 25) and an experimental (n = 25) group that used Scratch and OpenSim with the Scratch4SL palette, respectively to propose their solutions for the same problem-solving tasks via a simulation game. The study findings indicate substantial differences and important points of view about students' learning performance by assessing their computational problem-solving strategies. Students from the experimental group performed significantly better both in measures of problem-solving and algorithmic thinking. Mean scores on post-questionnaires from the experimental group revealed improvements higher than their control group counterparts in two aspects. First, students of the former group created more complete computational instructions with rules to be specified and delivered the learning goals. Second, students of the same group proposed and applied more correct computational concepts and practices in code. Finally, this study discusses the implications for designing learning experiences using OpenSim with Scratch4SL.