RESEARCH QUESTION: Do shorter but more frequent patrol visits to the same crime hot spots reduce daily crime and anti-social behaviour (ASB) totals more effectively than less frequent but longer patrols, if the total time that police are present each day is held roughly constant? DATA: GPS measures from patrol officer body-worn radios tracked the time each officer spent within seven geo-fenced crime “hot spots” of 150 × 150 m, summing the number of both individual officer-minutes and patrol-minutes (with one or more officers present simultaneously) per day per hot spot, as well as number of visits and minutes per visit. Activity reports were used to detect the simultaneous presence of more than one officer, yielding the key independent variables of number and length of patrol visits in which one or more police officers were present. The dependent variable was total reports of crime and anti-social behaviour (ASB) identified within the hot spot boundaries each day of the experiment. METHOD: All seven hot spots of crime and ASB were randomly allocated each day to one of two patrol duration conditions for a period of 100 days (43 “short” visit days and 57 “long” visit days) between June and November 2015, with patrol time measures reported back to officers on the number and length of patrols conducted daily. The long visit model required three visits daily of 15 min duration each; the short visit model required nine visits daily of 5 min each. On all days, a target of 45 min of total patrol time was required. RESULTS: Actual patrol delivery measured by GPS and activity reports produced a mean of just over 24 patrol-minutes (of one or more officers present) on “long” days and just under 26 min on “short” days, so that dosage was approximately held constant to test the independent effect of more or fewer visits. The treatment as delivered on “long” days was a mean of 2.5 visits averaging 9.6 min each; on “short” days, the same officers delivered a mean of 5 visits averaging 5.2 min each. The less-frequent long visit model was more effective than the more-frequent short visit model, with mean counts of crime and ASB incidents 19.51% lower on long visit days =0.697 incidents per day compared to 0.561 incidents per day on short visit days (d = −0.175; p = 0.018). CONCLUSION: Controlling for the total patrol time spent at a hot spot each day, the difference between 2.5 longer visits and 5 shorter visits causes about 20% less crime when longer visits are delivered. These findings of the deterrent effect of increasing patrol visit length by 85% are consistent with Koper’s (1995) correlational observation that longer units of 10–15 min duration appeared optimal in creating a residual deterrent effect at a hot spot immediately after police leave the vicinity. Although this study cannot distinguish between crime reductions immediately after vs. long after police have left the scene, this is the first experiment to randomly assign a substantial difference (twice as many) in the number of visits daily to a hot spot, with almost twice as much time per visit when fewer visits are made. The use of random assignment of two different patrol models with the same total time in the same seven geographic units gives great confidence that using that time in fewer visits of longer duration causes less crime and anti-social behaviour than more visits of shorter duration.