Add a reference probability of success to a metric-vs-x-variable plot — plot_baseline_success_proba_vs

Draws the baseline as points, optionally joined by a line. When selected_metric_uncertainty_lower/_upper name columns that carry a non-degenerate interval, the baseline also gets error bars.

The bounds are only informative where the power was approximated by Monte Carlo: compute_freq_power() returns an exact binomial interval when it simulates, but a degenerate c(power, power) when it has a closed form. Handing those degenerate bounds to geom_errorbar() would draw a zero-height bar - a bare cap tick - through every marker, so they are left off. Results that predate the bound columns are treated the same way, which keeps their figures rendering.

Usage

plot_baseline_success_proba_vs_xvar(
  plt,
  data,
  xvar_name,
  selected_metric_name,
  cap_size,
  markersize,
  join_points,
  label = NULL,
  selected_metric_uncertainty_lower = NULL,
  selected_metric_uncertainty_upper = NULL
)

Arguments

plt: The plot to add the baseline to.
data: The dataframe holding the baseline.
xvar_name: Name of the x-axis variable.
selected_metric_name: Name of the column holding the baseline.
cap_size: Width of the error bar caps.
markersize: Size of the points.
join_points: Whether to join the points with a line.
label: Legend label for the baseline.
selected_metric_uncertainty_lower: Name of the column holding the lower confidence bound, or NULL for no error bars.
selected_metric_uncertainty_upper: Name of the column holding the upper confidence bound, or NULL for no error bars.

Value

The plot, with the baseline added.