election_title: "Block approval voting — uncap the marks and the sweep reverses"
scenario_description: |-
  The same 9 voters as mmp_minority_sweep.yaml, with one rule changed: no cap on
  how many candidates a voter may mark.

  Under the 3-mark cap the two independent voters spend everything on Gus, Hugo
  and Iris, and never get to say that they prefer Pine to Oak. Uncapped, they say
  it: they approve their three independents AND the Pine slate. Nothing about
  their opinion changed — only how much of it the ballot could hold.

  That single extra sentence moves Dev, Enzo and Finn from 3 to 5 and hands Pine
  all three seats. Oak's 4 marks are unchanged; Oak simply stops being the
  largest number on the page.

  The lesson is not "approval is better" — block approval is still a majoritarian
  at-large method that can hand one bloc every seat, which is exactly what it does
  here. The lesson is that in this family the WINNER IS AN ARTIFACT OF THE CAP.
  Same voters, same opinions, three different answers: capped at 3 → Oak sweeps;
  uncapped → Pine sweeps; capped at 3 but run in two rounds → Pine sweeps
  (mmp_majority_block_runoff.yaml). For proportionality you need a different
  count, not a different mark limit — see STAR-PR or STV.

  Shrunk from the "Block approval voting" column of Wikipedia's block voting table
  (10,000 voters, 12 candidates).
voting_method: Approval_Multi_Winner
num_winners: 3
lot_numbers: [Alma, Bram, Cleo, Dev, Enzo, Finn, Gus, Hugo, Iris]
ballots: |-
  Alma,Bram,Cleo,Dev,Enzo,Finn,Gus,Hugo,Iris
  1,1,1,0,0,0,0,0,0   # Oak voter — approves the Oak slate only
  1,1,1,0,0,0,0,0,0   # Oak voter
  1,1,1,0,0,0,0,0,0   # Oak voter
  1,1,1,0,0,0,0,0,0   # Oak voter
  0,0,0,1,1,1,0,0,0   # Pine voter — approves the Pine slate only
  0,0,0,1,1,1,0,0,0   # Pine voter
  0,0,0,1,1,1,0,0,0   # Pine voter
  0,0,0,1,1,1,1,1,1   # independent voter — independents AND Pine
  0,0,0,1,1,1,1,1,1   # independent voter
expected_winners:
  - Dev
  - Enzo
  - Finn

# file: mmp_block_approval.yaml