election_title: "Reversal symmetry — RCV-IRV, reversed: A wins AGAIN (worst = best)"
scenario_description: |-
  The original election (reversal_irv_original) with EVERY voter's ballot reversed — each now
  expresses the exact opposite preference, as if trying to elect the WORST candidate. You would
  expect A (the original winner) to now finish last. Instead RCV-IRV elects A AGAIN (16-8): B has
  the fewest first-choices, is eliminated, and B's ballots flow to A. So IRV's "best" candidate and
  its "worst" candidate are the same — a reversal symmetry FAILURE. This is a real, IRV-specific
  defect (concede it). Caveats on ../README.md: the electorate is a Condorcet cycle, so there is no
  "correct" winner; STAR gives B then A (no winner=loser); source is Range-advocacy (lean disclosed).
paradoxes: [condorcet-cycle]
voting_method: RCV_IRV
num_winners: 1
ballots: |-
  9:A>C>B
  8:C>B>A
  7:B>A>C
expected_winners:
  - A
# file: reversal_irv_reversed.yaml