Accurate first-order transition points from finite-size data without power-law corrections

Traditional observables used to locate first-order phase-transition points from Monte Carlo simulations on finite lattices usually exhibit powerlike finite-size scaling behavior. For systems in a box of volume V with periodic boundary conditions, several definitions of finite volume (pseudo-) transition temperatures T₀(V) have recently been proposed which involve only exponential corrections with respect to T₀{≡}T₀({∞}). One of these approaches suggests also a different and efficient way to compute the latent heat. These propositions are tested for the temperature-driven first-order phase transitions in the two-dimensional q-state Potts model for q=5, 8, and 10 by means of Monte Carlo simulations, using the single-cluster update algorithm and histogram reweighting techniques.