Monte Carlo simulations for nuclear logging applications are considered to be highly demanding transport problems. In this paper, the implementation of weight-window variance reduction schemes in a 'manual' fashion to improve the efficiency of calculations for a neutron logging tool is presented. The simulations were conducted using MCNP to assess the sensitivity of neutron and gamma flux to the saturation of soil with fresh and salty water and oil. The weight-windows were optimised to enhance the precision in characteristic gamma-ray windows for carbon-oxygen logging and sigma logging. An analysis is under way to incorporate quartz sediments and other oil substances. Future tasks include the assessment of advanced Monte Carlo estimators and optimisation of capture gamma-ray sampling techniques to enhance the variance reduction in nuclear logging calculations.