NEWS

multigrain 1.0.0.9002

Renamed the multigrain_verbosityoption to multigrain.verbosity. The retrieval function (multigrain_verbosity()) remains unchanged.

The graphical shortcut used to score candidate graphs (in graph_optimise(), calc_power_pvals() and pruning) now pre-computes the local critical values for every intersection once per graph and looks them up in each simulated trial, instead of updating the graph after every rejection. This is faster for small and moderate numbers of hypotheses and gives the same rejections; the previous algorithm is still used for large numbers of hypotheses. There is no user-visible API change. The approach follows graph_test_shortcut_fast() in the graphicalMCP package (Xi and Brockmann, 2024).