The anticipation of events in time allows for action preparation and fast responses. This talk investigates how the probabilities of whether and when future events will occur shape human behavior. To predict when an event will occur, neural systems need to estimate elapsed time and event probability over time. According to a prominent hypothesis, neural systems fulfill these computational demands by estimating the hazard rate of events. Based on psychophysical experiments and computational modeling, we show that humans estimate a simpler variable: the event probability density function (Grabenhorst et al., Nat Commun, 2019). Neural systems are imprecise clocks, introducing uncertainty to time estimation. A common hypothesis poses that temporal uncertainty linearly increases with time itself (scalar property). This implies independence between time estimation and event probability. In contast, we demonstrate that event probability density modulates temporal uncertainty (Grabenhorst et al., Nat Commun, 2019). The probability of whether an event will occur is another fundamental variable in temporal anticipation. We show that this static occurrence probability has a highly dynamic effect on anticipation over time (Grabenhorst et al., PNAS, 2021). Finally, using magnetoencephalography, we show that the key variable in temporal anticipation – the event probability density – is represented in three neural signals, one in superior parietal lobule, one in inferior parietal lobule and posterior middle temporal gyrus and one in sensorimotor cortex, all prior to anticipated sensory cues (Grabenhorst et al., Nat Commun, 2025). Our results contribute to a mechanistic understanding of temporal anticipation – a fundamental process underlying many cognitive domains.