Skip to content

BV130-r2 — 6 candidates / 3 winners, Bloc STAR: a dead-rung lot tie

Method: Bloc STAR (multi-winner, majoritarian) · 3 seats · Expected winners: Dan, Ada, Eve · full count →

A real BetterVoting election (id 9ff9jk) built directly from this repo's tuned dead-rung scenario: 6 candidates, 3 seats, 4 ballots, and nothing scored above 4. Seat 1 is a perfect Ada/Dan tie that dies at every deterministic rung and falls to the lot. BetterVoting's random draw elected Dan, Ada, Eve; a pre-published order would elect Ada, Dan, Eve — same three winners, the tied pair's seats swapped.

Which BV130 is this? This is BV130-r2 (the retest, election 9ff9jk, marked Passed) — a different election from the original BV130, which is a 9-ballot "Someone I Like / Santa Claus / The Lesser Evil" case backing the star-server#731 reporting issue (tabs → numbered pages). See bv130_bloc_pagination_731.md for that one. This page is purely about the tie behavior; the outcome itself is valid (a random draw among a genuine tie), which is why the tracker marks it Passed.

Reference files: bv130r2_dead_rung_bloc.yaml (expected_winners: [Dan, Ada, Eve]) · frozen export bv130r2_dead_rung_bloc_bv_export.json (BV 9ff9jk). Backs sheet row BV130-r2.

The election

Bloc STAR, 6 candidates, 3 seats, 4 ballots — capped at 4, so no candidate ever holds a five:

Ada,Ben,Cara,Dan,Eve,Finn
4,1,2,3,2,0
3,2,1,4,2,1
4,2,1,4,3,2
4,2,1,4,3,1

Totals: Ada 15, Dan 15, Eve 10, Ben 7, Cara 5, Finn 4. Ada and Dan tie for the top, and because the ceiling is 4 their five-star counts are both 0 — the dead rung.

View 1 — BetterVoting

Elected Dan, Ada, Eve (nAbstentions: 0, nTallyVotes: 4 — all four ballots counted). The export's elected block confirms Ada and Dan each have score: 15 and fiveStarCount: 0, and its perm (the random draw) is:

[Finn, Ben, Dan, Eve, Ada, Cara]

Dan sits ahead of Ada in that draw, so the seat-1 Ada/Dan tie broke for Dan. But the result's top-level tieBreakType is reported as "none" — even though seat 1 was decided by a random draw. A reader of the summary can't tell a seat was settled by chance (#1379 family).

(Drop a BV screenshot into img/ as img/9ff9jk_result.png to add it here.)

View 2 — the LH report (reproducing BV's draw)

Pinning lot_numbers to BV's drawn sequence [Finn, Ben, Dan, Eve, Ada, Cara] reproduces Dan, Ada, Eve exactly. Every deterministic rung in seat 1 ties, the lot decides, and the engine flags it:

Abridged for the lesson — not verbatim engine output
[Score Distribution] (how many ballots gave each star rating)
                Score
Candidate  5  4  3  2  1  0  | Total   Avg
Ada        0  3  1  0  0  0  |    15   3.8
Ben        0  0  0  3  1  0  |     7   1.8
Cara       0  0  0  1  3  0  |     5   1.3
Dan        0  3  1  0  0  0  |    15   3.8
Eve        0  0  2  2  0  0  |    10   2.5
Finn       0  0  0  1  2  1  |     4   1.0

Round 1: Scoring Round
   Ada  -- 15 -- First place
   Dan  -- 15 -- Second place
 Ada and Dan advance.
Round 1: Automatic Runoff Round
   Ada  -- 1 -- Tied for first place
   Dan  -- 1 -- Tied for first place
   Equal Support -- 2              ← runoff tie
Round 1: First tiebreaker (highest score)
   Ada  -- 15 ;  Dan  -- 15        ← score tie
Round 1: Second tiebreaker (most 5s)
   Ada  -- 0 ;  Dan  -- 0          ← five-star DEAD (nothing reaches 5)
[Tiebreaker: Lot Number Priority]
  Tie among: ['Ada', 'Dan']  →  Resolved: ['Dan']   (lot [Finn, Ben, Dan, Eve, Ada, Cara])

[Lot-decided tie — rare]
  ⚠ The ballots did not break this tie ... the LOT order chose the winner.

Round 2: Scoring Round — Ada (15) and Eve (10) advance; Ada wins the runoff 100%.
Round 3: Scoring Round — Eve (10) and Ben (7) advance; Eve wins.

Winners — Bloc STAR Voting Method (3 winners)
 Dan
 Ada
 Eve
Full audit copy: _main_tabulated/bv130r2_dead_rung_bloc_tabulated.txt.

Two findings

  1. Non-reproducible (cf. #1063 / #1417). With a published lot order [Ada, Ben, Cara, Dan, Eve, Finn] the seat-1 tie breaks for Ada, giving Ada, Dan, Eve — the same three winners, but seats 1 and 2 swap between the tied pair. Which candidate takes which seat is decided purely by the draw; BV's random happened to pick Dan.
  2. Reporting mislabel. The result's top-level tieBreakType: "none" contradicts a seat that was decided by the lot (cf. #1379). The summary should surface that a seat was lot-decided.