Add error estimate to sensitivity/discovery potential

[robertdstein]

Is your feature request related to a problem? Please describe.
I'm always frustrated when you calculate a sensitivity, but have to judge by eye how reliable the estimate is.

Describe the solution you'd like
Use sqrt(n) as an assumption for the error, curve fit, and propagate through to get a range of sensitivities. Error on thresholds themselves would be more difficult, but possible.

Describe alternatives you've considered
Leaving things as they are. Not wrong, but less informative.