election_title: "BV2209 — Burial in Ranked Robin (2/2): fifteen voters rank the leader last, and it pays"
bv_test_id: BV2209
bv_election_id: fxhw6g
bv_results_url: https://bettervoting.com/fxhw6g/results
scenario_description: >-
  The burial. Same 42 voters as part 1 (bv2208_7q6by8_burial_sincere.yaml),
  except the 15 Amber-first voters now rank Amber>Coral>Diamond>Beryl —
  burying Beryl, their honest SECOND choice, below two gems they like less.
  The mechanics: burial's reach is exactly the buriers' own weight inside
  the victim's coalitions. Beryl's 33-9 over Coral and 27-15 over Diamond
  both contained the buriers' 15 ballots, so withdrawing them flips both
  (Coral 24-18, Diamond 30-12). Her 27-15 over Amber contained none of them
  (they already ranked Amber first) — untouchable, which is why Amber can
  never beat Beryl DIRECTLY and must win through the record instead. The
  round-robin becomes a cycle with Amber and Coral tied on top at 2-1, and
  Amber takes the tie on EVERY metric: total pairwise margin +12 vs 0 (LH's
  rung), the direct head-to-head 27-15 (BV's rung), first choices 15 vs 9.
  The buriers turned Beryl's win into Amber's — and left fingerprints: a
  cycle where sincere ballots showed none. Burial needs a large coordinated
  bloc sitting inside the leader's own majorities, good polling, and a lie;
  the printed round-robin table is where you'd catch it.
  (Contrast BV2142: a THREE-way RR tie is random on BV — this pair stays
  freezable because a 2-way tie resolves by head-to-head, deterministically.)
  Triple-checked: LH native (margin rung), pref_voting Copeland-leader set
  {Amber, Coral}, BetterVoting live (Amber, tieBreakType none).
  Live results: https://bettervoting.com/fxhw6g/results
voting_method: RankedRobin
num_winners: 1
ballots: |-
  15:Amber>Coral>Diamond>Beryl
  12:Beryl>Amber>Diamond>Coral
  9:Coral>Diamond>Beryl>Amber
  6:Diamond>Beryl>Coral>Amber
expected_winners:
  - Amber

# file: bv2209_fxhw6g_burial_pays.yaml