Add (binomial) confidence intervals for coverage

[maxbiostat]

Table tab:coverage would benefit from some bounds on 'acceptable' bounds so one can readily check whether things are working.