If you get burgled in Leicestershire, police will only come and investigate if you live in an even numbered house. Seriously.
The Telegraph reports that it is all part of a three month trial to determine whether it will impact upon crime rates or public satisfaction. As part of the trial, police would only send forensics officers to even numbered homes that report attempted burglaries. Any crime that had either a potentially vulnerable victim or was suspected as being part of a related crime spree would continue to be investigated, though.
And here's the slightly counter-intuitive thing: It turns out to have been a really good idea. According to the Telegraph piece, there was no noticeable impact and of 1172 forensics visits only 33 suspects were identified. In other words, it seems that visits until now have been a pretty massive waste of resources - and this test has proved it. There is now talk of rolling out the same "evens only" policy to other areas to test it further afield.
Randomised control trials in public policy like this have been advocated by academics for some time, over a wide number of fields but are always likely to be controversial. One particularly contentious area is control trials over the distribution of humanitarian aid, as such trials could help experts figure out what the most effective interventions are (and thus the most effective way of spending our cash), but just as this police trial has literally stopped crimes from being thoroughly investigated, such a humanitarian aid trial would (to put it crudely) be like deciding who gets food and who doesn't. [The Telegraph]