Myocardial fibers of the left ventricle (LV) play a pivotal role in electrical conduction, mechanical contraction, and numerous clinical malfunctions. While the general fiber orientation of the LV has been revealed through histological analysis and magnetic resonance diffusion tensor imaging, its impact on LV deformation remains largely unknown. In this paper, we adopt an idealized hollow semi-ellipsoid LV model, allowing for adjustable fiber orientations using a widely-accepted rule-based method. Simulations are conducted using a robustly coupled excitation-contraction nonlinear finite element algorithm. Our primary focus is on exploring the orientation angle of regularly-distributed fibers and the proportion of chaotic fibers, whose orientation angles are randomly assigned, on the end-systolic volume and ejection fraction of the LV. By employing this model, we successfully recreate the changes in LV volume over a cardiac cycle and capture the typical twisting motion observed in clinical practice. Furthermore, our findings reveal that when myocardial fibers are regularly distributed and the orientation angle increases, the ejection fraction of the LV decreases along with an increase in end-systolic volume, indicating a decline in LV contractility. Additionally, both the proportion and spatial distribution of chaotic fibers within the LV influence its contractility. Specifically, an LV with a higher proportion of chaotic fibers in the basal area exhibits weaker contractility. These results provide deeper insights into the quantitative influence of myocardial fibers on LV contractility and failure, offering valuable information for further research and clinical applications.