Gates' bidding model allows anticipating the probability of submitting the lowest bid in a future auction. Despite its relative simplicity, this classical model has been shown to outperform many other bidding models in real auction settings. However, Gates' model is accurate if, and only if, bidders' bid probability distributions are from the proportional hazards family. Unfortunately, checking this assumption in practice is difficult ex-ante (before the auction takes place) due to limited access to similar previous auctions' information. In this paper we propose an approach to quantitatively measure the tenability of the proportional hazards assumption in real auction settings. By resorting to Monte Carlo simulation, we develop a method to measure the consistency of the pairwise probability matrix that stores the probabilities of every bidder individually underbidding each other. Application of our method will allow bidding decisionmakers to assess Gates' forecasts reliability for upcoming auctions.