The idea of evidence–based medicine is now deeply ingrained in public expectations of how our health services operate. Evidence–based policing though is still relatively new, which is strange when we are equally concerned with preserving life and promoting safety. The symptoms police deal with are no less important than those presented to the NHS every day, although how we respond to them is very different. If there is a spike in thefts from a city centre we immediately diagnose the problem: it’s those new smartphones, it’s organised crime, it’s insurance scams, it’s because of welfare cuts, it’s proximity to that hostel. Well, which one is it? Then, too often, we go straight to a treatment: shut down the hostel, target a certain group of people, put an extra 100 officers on duty. Some of those actions may work and some may cause further harm. An evidence–based approach tackles hypotheses systematically, making sure it is underpinned by accurate data (both qualitative and quantitative). It then looks at what has been proved to work in real world contexts. In the world of crime, like education, medicine and other fields, a before–and–after analysis is far too weak. Too many other factors operate on people to make them ill, commit crime or achieve bad grades. Test and control groups are easy but rarely used. Randomised control trials are rare.