election_title: "Reversal symmetry — STAR, original: B (STAR does not winner=loser here)"
scenario_description: |-
  The same 24-voter cycle scored 0-5 (rank->score 5/3/0). STAR elects B (scoring round B 69,
  runoff 17-7). The companion reversal_star_reversed reverses every preference and STAR elects
  A — a DIFFERENT winner, so STAR does not exhibit the winner=loser failure RCV-IRV shows on this
  example (see ../README.md). Honest scope note: this shows STAR AVOIDS the failure on these
  ballots; it is not a claim that STAR satisfies the reversal-symmetry criterion in general (STAR
  is a hybrid — additive scoring round + pairwise runoff — and the rank->score mapping is a modeling
  choice). The electorate is a Condorcet cycle, so no winner is "correct."
paradoxes: [condorcet-cycle]
voting_method: STAR
num_winners: 1
ballots: |-
  Count:A,B,C
  9:0,5,3
  8:5,3,0
  7:3,0,5
expected_winners:
  - B

# file: reversal_star_original.yaml