Recent technological advances in high-resolution imaging and artificial modulation of genetic functions at different times and regions have enabled direct observations of the formation and elimination of abnormal cell populations. A recent trend in cell competition research is the incorporation of cell mechanics. In different tissues and species, abnormal cells developing in epithelial tissues are mechanically eliminated by cell contraction via actomyosin accumulation at the interface between normal and abnormal cells. This mechanical cell elimination process has attracted attention as a potential universal defense mechanism. Here, we theoretically examined the conditions for mechanical elimination of growing abnormal cell populations. Simulations and mathematical analyses using a vertex dynamics model revealed two types of critical cell density associated with mechanical elimination of abnormal cell clusters. One is a subtype of homeostatic density, in which the frequencies of spontaneous mechanical cell elimination and proliferation are balanced, even if no explicit dependence of proliferation or apoptosis on the cell density is assumed. This density is related to the mechanical stability of a single cell. The other is density related to mechanical stability as a cell population under external pressure. Both density types are determined by tissue mechanical properties. In solid tissues, the former type is reached first as the intensity of interfacial contraction increases, and it functions as a critical density. On the other hand, the latter type becomes critical when tissues are highly fluid. The derived analytical solution explicitly reveals the dependence of critical contractile force and density on different parameters. We also found a negative correlation between the proliferation rate of abnormal cells and the likelihood of the abnormal cell population expanding by escaping elimination. This is counterintuitive because in the context of cell competition, fast-growing cell populations generally win. These findings provide new insight into, and interpretation of, the results from experimental studies.