Showing confidence estimates

[konnov]

Champion: @angbrav

This discussion presents a potential feature that could help @angbrav in the partnership with Heliax. The following discussion is presenting @angbrav's story, written in my own words.

Given that am developing a protocol in a partnership (for several months), I am writing a Quint specification of this protocol.

When I run the simulator quint run --max-samples=N.

Then I would like to know how well the simulator has covered the specification. In particular, when I increase the number of runs to 10 * N, how much would it help me to achieve better confidence?

This story can be implemented via the following features:

• Implement line coverage in the simulator #1068 (3-5 days)
• Estimate the number of visited states in the simulator #1069 (1-2 weeks)