Short infrequent police stops maximize crime-free time at street segments: revisiting the Koper curve to establish optimal police dosage

Establishing optimal police dosage is central to evidence-based policing. The current consensus is that 15-minute patrols represent the optimal level of such dosage. This study revisits this central assumption of police research and practice, by examining the relationship between the duration and frequency of police patrol and the timing of subsequent crime at the street segment level. Drawing on 21,581 GPS-tracked police stops and 45,987 citizen-reported crimes across 25,844 street segments in Antwerp, Belgium, during 2019, we employ accelerated-failure time survival models to estimate time-to-subsequent-crime. Our findings challenge the 15-minute standard: Shorter, infrequent police visits—particularly those lasting 6–15 minutes and occurring once every 24 hours—are associated with longer crime-free intervals across all street types, including hot spots and the most crime-prone ‘hottest spots’. In contrast, prolonged or more frequent patrols yield diminishing returns. These results suggest that tailoring patrol strategies to local crime concentrations and embracing variability in dosage may enhance deterrence while conserving police resources. We argue for a re-evaluation of the Koper Curve and call for more granular, context-sensitive approaches to patrol planning.