Sequential foreperiod effects in temporal preparation are typically asymmetric such that a previous experience of preparation has a strong impact on participants’ responses to a forthcoming target stimulus presented at the current short rather than at the current long foreperiod. The trace-conditioning model explains this asymmetry by an automatic process of trace-conditioning, which is sufficient and independent from a strategic process of conditional probability monitoring considered by the dual-process model. The present study contrasted trace-conditioning and dual-process models in three experiments that employed a non-aging distribution to keep conditional probability of target occurrence constant across foreperiod durations. In Experiment 1 (no catch trials), the typical pattern of asymmetric sequential effects was replicated, whereas in Experiments 2 and 3 (25% and 50% of catch trials, respectively) the results showed shorter RTs when previous and current foreperiods were repeated rather than alternated for both current short and long foreperiods. These results are discussed in relation to the two most influential models of sequential effects and to a novel account based on repetition priming.