As a next step I would love to see an exploration of a legal system where a change of a law or regulation is backed by both data (data and methodology directly referenced by the law) and a description of the expected impact. Something such as:

> By changing law A related to B, we expect the increase of C to be at least D in the next E months.

If not achieved, the change is reversed/reduced. Hopefully that would allow experimentation without taking the risk of creating a system too bad in case the implementation or policy isn’t good enough.

Just dreaming here :)