We propose a novel approach to the modelling of bidding behavior in pay-your-bid auctions that builds on the presumption that bidders are mostly concerned with losing an auction if they happen to have the highest signal. Our models assume risk neutrality, no profit maximization and no belief about competitors’ behavior. They may entail overbidding or Symmetric Bayes-Nash Equilibrium bidding and we discuss conditions for the revenue equivalence of standard pay-your-bid auctions to hold. We fit the models to the data of first-price auction experiments and find that they do at least as well as the Bayes-Nash equilibrium benchmark model in organizing the observed behavior. Assuming probability misperception or impulse weighting (when relevant) improves their goodness-of-fit and leads to very similar revenue predictions. An analysis of individuals’ heterogeneous behavioral traits suggests that impulse weighting is a more consistent rationale for the observed behavior than a power form of probability misperception.