Axis breaks that always show the nominal type-I error

pretty() picks breaks that ignore the nominal TIE, so forcing the nominal value in alongside them leaves two labels sitting on top of each other. check.overlap resolves that collision the wrong way round: it draws labels leftmost-first and discards the nominal tick, which is the one the plot is read against. Drop the neighbouring pretty breaks instead, so the nominal value is the only label in its neighbourhood.

Usage

nominal_tie_breaks(nominal_tie)

Arguments

nominal_tie

The nominal type-I error rate, or NULL when the scale has no nominal value to mark.

Value

A function of the scale limits returning a sorted break vector.