The defect this harness exists to quantify: comparing only the layer-name sequence counts two voters as agreeing whenever they pick the same layers, even when every argument differs. This reports the gap between the `layer_chain` and `full` agreement rates - i.e. the share of cases the old [evals_layers_label()] comparison would have scored as agreement but which differ semantically.
The result inherits every caveat of [run_agreement()]: it is computed pre-execution over voters reading the same spec text from a self-scoring regression corpus, so it quantifies overstated translation consistency, not correctness and not independent double programming.
Value
A list with `layer_chain`, `full`, `overstatement` (the difference) and `cases` (the ids counted as agreeing only at `layer_chain`).
Examples
res <- run_agreement(list(a = "rules", b = "rules"), dataset = "ADSL")
agreement_overstatement(res)
#> $layer_chain
#> [1] 1
#>
#> $full
#> [1] 1
#>
#> $overstatement
#> [1] 0
#>
#> $cases
#> character(0)
#>