The pin-based pointwise energy slowing-down method (PSM) has been refined through eliminating the approximation for using the pre-tabulated collision probability during the slowing-down calculation. A collision probability table is generated by assuming that material composition and temperature are constant in the fuel pellet using the collision probability method (CPM). Refined PSM (PSM-CPM), which calculates the collision probability in the isolated fuel pellet during the slowing-down calculation using CPM, can consider nonuniform material and temperature distribution. For the methods, the extensive comparative analysis is performed with problems representing various possible conditions in a light water reactor (LWR) design. Conditions are categorized with the geometry, material distribution, temperature profile in the fuel pellet, and burnup. With test problems, PSMs (PSM and PSM-CPM) have been compared with conventional methods based on the equivalence theory. With overall calculation results, PSMs show the accuracy in the eigenvalue with differences in the order of 100 pcm compared to the reference results. There was no noticeable difference in the multigroup cross sections, reaction rates, and pin power distributions. However, PSM-CPM maintains the accuracy in the calculation of the fuel temperature coefficient under the condition with 200% power and nonuniform temperature distribution in the fuel pellet. PSM shows the difference in the eigenvalue in the order of 2,000 pcm for the fictitious pin-cell problem with highly steep temperature profiles and material compositions, but PSM-CPM shows the difference in the eigenvalue within 100 pcm.