Execute simulations for marginal operating characteristics

Calculates marginal operating characteristics using Monte Carlo simulation with Maurer-Bretz graphical procedure. Generates 3D rejection array and computes comprehensive power metrics including individual hypothesis power, subset power, and expected analysis timing.

Usage

exec_sims(
  analyses,
  hypotheses,
  graph,
  method = "z",
  sims = 1000,
  alpha = 0.025
)

Arguments

analyses: Processed analyses data frame
hypotheses: Processed hypotheses data frame
graph: Graph object with transition matrix and weights
method: Method for test statistics (default "z")
sims: Number of simulations
alpha: Overall Type I error rate

Value

List with marginal_oc containing:

oc_at_analyses: Power metrics by analysis (individual, any, all)
oc_across_analyses: Expected analysis index and expected time for first rejection

Details

Workflow: 1. Collapse hypotheses by endpoint, extract distribution parameters 2. Generate MVN distribution for correlated test statistics 3. Process rejection rules and alpha spending thresholds 4. Simulate p-values from MVN, apply Maurer-Bretz to each simulation 5. Build 3D rejection array [J hypotheses × K analyses × S simulations] 6. Compute power metrics via get_sim_oc()