Yes. It particularly shows an example of how bad incidents (e.g. bugs, serious offences) tend to be underreported or have their criticality misrepresented. I feel that whenever possible, these metrics should be used for strategic guidance, but the primary metrics for people in the field to follow should be positive ones—things that the team agrees will likely affect the negative metrics in a desirable manner: number of test cases, reviewed commit counts, emergency response times or serious lawbreakers caught in the act.

But many of the things that affect the final outcome may not be particularly measurable. Without expert knowledge of law enforcement I think that may be the problem with evaluating police officers' work. For example, conversations with citizens may prove to aid crime reduction and confidence in the police, but how do you quantify them in terms of nature and value?

It goes back to being careful what you measure, because it can be gamed. In this case, it was gamed to very unfortunate consequences.