Court delays are a frequent concern, yet what explains court case duration remains incompletely understood. We study the time to court case resolution by drawing on a detailed case-level dataset of civil suits filed at a major Belgian court. We utilize the competing risks regression framework to address the typically neglected heterogeneity in the modes of court case resolution and examine the role of a wide range of both time-invariant and time-varying covariates. Controlling for judge fixed effects, we find substantial disparities in the effect of party and case characteristics on the time to settlement versus trial judgment. Exploiting the de facto random assignment of cases to serving judges within the court's chambers, we further find that judge characteristics matter for time to trial judgment, but not for time to settlement. Modeling heterogeneity in the modes of court case resolution is therefore central to understanding of court case durations.