Comparative judgment biases—wherein a majority of people report being above‐ or below‐average in their abilities, traits, or future events—are a robust phenomenon in psychology. A recent explanation for these biases has focused on people's awareness that many comparative judgment domains form skewed distributions, and, hence, a majority of people can feasibly be above or below average. Indeed, this prior research found that comparative biases for abilities emerged more for skewed (vs. normal) distributions. In the current research, we attempted to (i) conceptually replicate this finding in a comparative likelihood context and (ii) provide evidence of an alternative explanation for the prior results. Replicating prior research, three correlational studies and one experimental study found that event skewness was related to direct comparative likelihood judgments for health events, such that comparative optimism emerged more for events judged or manipulated to come from positively skewed distributions than from negatively skewed distributions. However, event skewness was unrelated to indirect comparisons (absolute self minus absolute other). Moreover, consistent with an egocentric‐processes account, absolute self‐judgments were more predictive of direct comparisons than were absolute other judgments and showed the same association with event skewness as direct comparisons. Implications for explaining and interpreting comparative judgment biases are discussed. Copyright © 2016 John Wiley & Sons, Ltd.