Skip to content

BV2105 — Favorite ice cream (Bloc STAR): a partial ballot mis-filed as an abstention

Method: Bloc STAR (multi-winner, majoritarian) · 2 seats · Expected winners: Chocolate, Strawberry · full count →

BetterVoting test BV2105 (r4dqvd, "Favorite ice cream (Bloc STAR) - without end date"). Bloc STAR, 3 flavors, 2 seats, 4 ballots. This page reads the export frozen on 2025-10-31. In that count the winners were right (Chocolate, Strawberry) and the seat-2 score tiebreak matched LH exactly — but the tally was wrong: BetterVoting filed a real partial ballot as an abstention, reporting nTallyVotes 2 / nAbstentions 2 where the truth is 3 tallied + 1 abstention. Still reproducing on a fresh election a year later — see BV2105-r2.

Citation correction (2026-08-04). This page used to credit the miscount to bettervoting#1056 and call it fixed. Both were wrong. #1056 is a different defect on the same demo election — a 401 blocking JSON/CSV download and Race Details, introduced by the Editable Ballots work (#979) and correctly closed via #1058. The two share only the BV2105 test-document name. The counting defect below is a separate bug, filed 2026-08-04 as #1478.

▶ Live on BetterVoting: vote · results ↗ (election r4dqvd).

Reference file: bv2105_r4dqvd_ice_cream_bloc.yaml (expected_winners: [Chocolate, Strawberry]). Frozen export: bv2105_r4dqvd_ice_cream_bloc_bv_export.json. Backs sheet row BV2105.

The election

Bloc STAR, 3 flavors, 2 seats, 4 ballots:

Vanilla,Chocolate,Strawberry
5,5,5     an all-5s ballot (loves everything)
-,-,-     fully blank — a TRUE abstention
1,-,-     Vanilla=1, the rest blank — a REAL (partial) vote
2,5,4     a full ballot

Chocolate takes seat 1. Seat 2 is a Vanilla/Strawberry runoff tie (1–1), broken by score — Strawberry 9 > Vanilla 8 — so BetterVoting's tieBreakType: score and LH agree. Winners: Chocolate, Strawberry.

The regression (counting, not tabulation)

BetterVoting's summaryData reports nTallyVotes: 2, nAbstentions: 2 — it counts only the two full ballots and files both the blank ballot and the partial 1,-,- ballot as abstentions. You can see the partial ballot was dropped from the tally, not just mislabelled: BV's per-candidate score is an average over 2 ballots, so Vanilla reads (5+2)/2 = 3 — the 1 from ballot 3 never entered the sum.

The LH engine counts it correctly: 4 ballots, 1 abstention (only the fully-blank row), Vanilla total 8 (5 + 1 + 2). Same winners here — the discarded ballot only helped Vanilla, the seat-2 loser — so this is a reporting/counting regression rather than a wrong result. But a dropped cast ballot can flip a closer election, which is why it earns a guard.

This is the opposite-direction sibling of BV15 / bettervoting#740: #740 drops abstentions from the displayed turnout; BV2105 mis-classifies a real partial ballot as an abstention. Both are abstention-accounting defects.

The behavior is unchanged a year on. The re-check ran on 2026-08-04: the same four ballots, cast fresh so today's tabulator counts them, come back nTallyVotes 2 / nAbstentions 2 — identical to the numbers frozen here. That re-run is its own case, BV2105-r2 (w3vvff), which also explains why re-fetching this election couldn't have answered the question (a closed election may serve a stored 2025 result). So this page is the 2025 baseline; BV2105-r2 is the live evidence.

The likely root is #884's policy that a ballot whose marks are all equal counts as an abstention: a ballot bearing one mark is trivially all-equal, so it falls through the same test. That reading would make this intended behavior rather than a coding slip — which is precisely why it's worth arguing on the merits. A voter who scores Vanilla 1 and leaves the rest blank has expressed a preference, because BetterVoting's own tally treats those blanks as 0.

The LH report (the correct count)

Abridged for the lesson — not verbatim engine output
--- Bloc STAR Voting Method (2 winners) ---
 Tabulating 4 ballots to fill 2 seats. Note: 1 of 4 ballots is marked as an abstention.
Vanilla,Chocolate,Strawberry
      5,        5,         5
      -,        -,         -
      1,        -,         -
      2,        5,         4

[Score Distribution] (how many ballots gave each star rating)
                 Score
Candidate   5  4  3  2  1  0  Abs  | Total  Avg all  Avg rated
Vanilla     1  0  0  1  1  0    1  |     8      2.0        2.7
Chocolate   2  0  0  0  0  0    2  |    10      2.5        5.0
Strawberry  1  1  0  0  0  0    2  |     9      2.3        4.5

Round 1: Chocolate (10) & Strawberry (9) advance → Chocolate wins the runoff (1 vs 0, 3 Equal Support).
Round 2: Strawberry (9) & Vanilla (8) advance → runoff ties 1–1 → score tiebreak → Strawberry (9 > 8).

Winners — Bloc STAR Voting Method (2 winners)
 Chocolate
 Strawberry
Full audit copy: _main_tabulated/bv2105_r4dqvd_ice_cream_bloc_tabulated.txt.

The two average columns are what make the disagreement legible, because the whole defect is a denominator. Vanilla's total of 8 is not in dispute; what to divide it by is. LH offers the two defensible readings — 2.0 over all 4 ballots cast (a blank scores 0) and 2.7 over the 3 that scored her — while BetterVoting divides by 2, a denominator that matches neither, because it has already discarded the partial ballot whose 1 is inside that total. → how the two columns differ.

LH ↔ BetterVoting

Quantity BetterVoting summaryData LH engine Agree?
Winners Chocolate, Strawberry Chocolate, Strawberry ✓
Seat-2 tiebreak score (Strawberry) score (Strawberry 9 > Vanilla 8) ✓
Ballots tallied nTallyVotes 2 3 ✗
Abstentions nAbstentions 2 1 (blank only) ✗
Vanilla total 3 (avg over 2 ballots) 8 (5 + 1 + 2) ✗

The winner path matches; the count is where the bug shows — BV drops the 1,-,- ballot as an abstention. Still true today, on a fresh election: BV2105-r2.