Policing strategies and practices increasingly incorporate geographic information systems (GIS) with advanced data analytics. The practice of hot spots policing uses this technology to identify small geographic areas where crime is concentrated. This enables agencies to target their limited resources on the geographic areas of greatest need. More recently, however, analytical techniques have been applied to forecast crime. These forecasts may focus on places where crime occurs or on persons (perpetrators, victims, or witnesses). Together with relevant interventions, this practice has been called "predictive policing." The Shreveport Police Department (SPD) compared a predictive policing model focused on forecasting the likelihood of property crime occurring within block-sized areas, with a hot spots policing approach in a randomized, controlled experiment. The predictive policing approach targeted special operations to locations where property crime was forecast to occur. The hot spots approach, on the other hand, targeted special operations to locations where clusters of property crimes had already occurred. The independent evaluation of this comparative research found no significant differences in levels of property crime between the two strategies. The evaluation points to implementation failures, low statistical power, and problems with the underlying theory of the program as potential explanations for the lack of difference in property crime levels between the two strategies. In addition, the evaluation notes the possibility that the forecast did not add enough new information over the conventional hot spots method to make a difference in policing activity. The need for further research is discussed.