BV2142 — clone independence (1/2): a no-Condorcet cycle, and where LH and BV part ways¶

Method: Ranked Robin (RCV-RR / Copeland) · 1 seat · Expected winner: A · full count →

▶ Live on BetterVoting: vote · results ↗ (election 4gfwdq).

Part 1 of the electowiki clone-independence pair (the "before"). It sets up the teaming attack in BV2143, and on its own it already shows LH and BetterVoting breaking a tie differently. Full lesson: rr_clone_independence.md.

The setup¶

33 voters. A, B, C sit in a cycle (A beats B, B beats C, C beats A) — no Condorcet winner — and all three tie at 4 wins.

12:A>B>C>D>E>F
11:B>C>A>D>E>F
10:C>A>B>D>E>F

View 1 — BetterVoting: random¶

BV finds the 3-way tie and, with no margin rung for three tied candidates, picks at random:

tied: [C, B, A] — all copelandScore: 4
tieBreakType: random
elected: [C] (this draw); log: "C picked in random tie-breaker, more robust tiebreaker not yet implemented."

BetterVoting result page for 4gfwdq: headline 'Tied! — C won after tiebreaker', 33 voters, Voting Method: Ranked Robin. Head-to-head wins bars — C, B and A all at 80% (4 of 5 matchups), D 40%, E 20%, F 0%.

View 2 — LH: margin, then lot¶

LH ranks the tie by total win margin — and A and B tie there too (+101 each) while C is lower (+95). So LH drops C and settles A-vs-B by lot (pinned to A):

    #  Candidate  W–L–T  Copeland  Margin  Beats
    1  A          4–1–0         4    +101  B, D, E, F
    2  B          4–1–0         4    +101  C, D, E, F
    3  C          4–1–0         4     +95  A, D, E, F
Winner — Ranked Robin (RCV-RR): A

The divergence is real, not just a different coin: LH's margin rung would never elect C (lowest margin), yet BV's random draw did. Frozen export: bv2142_4gfwdq_clone_cycle_pre_bv_export.json. Full LH detail: _tabulated.

Next¶

BV2143 — teaming: the A-faction runs clones and the two engines end up with opposite winners.