I found Reed et al's (Reference Reed, Russell and Xenitidis2004) study fascinating, as it demonstrates the apparently random nature of a forensic label in our patients. It is clearly not to do with risk. I am confused by some of the results. The whole gist of the argument is that the offender group is less violent than their non-offender counterparts. However, it is stated that in the offender group the challenging behaviour diminishes from 0.79 incidents per week to 0.36 and that for the non-offender group from 0.23 to 0.11. This is challenging behaviour generally but this suggests that those in the offender group exhibit greater challenging behaviour throughout their stay than those in the non-offender group. Table 2 states the opposite. I would be interested to see how this inconsistency can be explained.