Studies investigating the perception of audiovisual simultaneity often start from the idea that the brain must compensate for the different arrival times and processing delays of auditory and visual events in order to determine whether two events were physically synchronous. We explored a very different notion. Given that our audiovisual experiences are extended in time, we asked whether the brain might use a strategy of temporal pattern matching to integrate audiovisual events. Visual and auditory event streams with the same pattern over time (like the lip movements and speech sounds of a person talking) are unlikely to match by chance, and therefore are likely to have a common cause. Such temporal pattern matching could be a reliable cue for multisensory integration that does not rely on precise sensory timing or stimulus-dependent compensation processes. In this study, we asked participants to match irregular streams of simple auditory and visual stimuli in a 2-alternative forced choice task. We showed that people could use common temporal pattern as an audiovisual correspondence cue over a wide integration window. The effectiveness of this cue depended on the temporal proximity between corresponding auditory and visual streams, but it was not specially enhanced when the streams were physically synchronous. The data suggested that correspondence perception followed the pattern of events over time, and not the absolute passage of time. Further, we found that the structure of the temporal pattern was important for correspondence detection. Counterintuitively, predictable, rhythmic event streams were harder to match crossmodally than stochastic, irregular streams. These results can be considered in terms of the information carried by different kinds of temporal structures.