The score leader can win no seat¶

One line: Bloc STAR is not "the top N by points." Every seat is settled by the automatic runoff, so seats go to whoever is preferred head-to-head — and a broadly acceptable compromise candidate can lead every scoring round by a wide margin, reach every runoff, lose every one of them narrowly, and end the night with nothing.

→ The single-winner version of the same step: runoff reversal · the mechanics: Bloc STAR · not to be confused with the majority sweep

Level: 201 → 301 · deep dive

The claim people actually make¶

Ask around and Bloc STAR gets described as "score everyone, the top N win." Half the repo's own shorthand says it that way, and for a lot of elections it happens to be true. It is still the wrong model, and it fails exactly where multi-winner elections get interesting.

The correct model is the one in the method page: each seat runs the whole STAR count, scoring round and runoff. The scoring round only picks two finalists. The runoff picks the winner. So the score order is a shortlist, not a result — and a candidate can top that shortlist forever without ever converting it.

BV1835 — 100 voters, 4 seats, and the leader takes nothing¶

Five candidates, four seats, so exactly one candidate loses. The electorate is two mirror-image camps of 49 that share no candidates at all, plus a 2-voter swing bloc:

Count × Ava,Bianca,Cedric,Deegan,Eli
   25 ×   3,     5,     4,     0,  0
   25 ×   3,     0,     0,     5,  4
   24 ×   3,     4,     5,     0,  0
   24 ×   3,     0,     0,     4,  5
    2 ×   0,     5,     4,     3,  2

Ava is the compromise. 98 of the 100 voters score her a 3 — above the pair they zeroed out, below their own two. Nobody's favorite, almost nobody's enemy. She leads the scoring round by 63 points:

[Divergence from STAR]
  STAR     = Bianca
  Approval = Ava   (differs from STAR)

[Runoff Reversal]
 - Score Round Winner(s) = (Ava)
 - Runoff Round Winner   = (Bianca)
  Candidate Ava earned the highest total score, but
  Candidate Bianca won the automatic runoff — not a malfunction,
  STAR working as designed: the runoff elects the finalist preferred
  by the majority (of voters with a preference).

--- Bloc STAR Voting Method (4 winners) ---

[Bloc STAR]
 Tabulating 100 ballots to fill 4 seats.
Count × Ava,Bianca,Cedric,Deegan,Eli
   25 ×   3,     5,     4,     0,  0
   25 ×   3,     0,     0,     5,  4
   24 ×   3,     4,     5,     0,  0
   24 ×   3,     0,     0,     4,  5
    2 ×   0,     5,     4,     3,  2

[Bloc STAR: Round 1: Scoring Round]
 The two highest-scoring candidates advance to the next round.
   Ava           -- 294 -- First place
   Bianca        -- 231 -- Second place
   Cedric        -- 228
   Deegan        -- 227
   Eli           -- 224
 Ava and Bianca advance.

[Bloc STAR: Round 1: Automatic Runoff Round]
 The candidate preferred in the most head-to-head matchups wins.
   Bianca        -- 51 -- First place
   Ava           -- 49
   Equal Support --  0
 Bianca wins.
   Runoff math:
     100  ballots cast
   −   0  Equal Support (no preference between the two finalists)
     ───
     100  voters with a preference  (majority = 51)
           Bianca 51 (51%)  ·  Ava 49 (49%)

──────────────────────────────────────────────────

[Bloc STAR: Round 2: Scoring Round]
 The two highest-scoring candidates advance to the next round.
   Ava           -- 294 -- First place
   Cedric        -- 228 -- Second place
   Deegan        -- 227
   Eli           -- 224
 Ava and Cedric advance.

[Bloc STAR: Round 2: Automatic Runoff Round]
 The candidate preferred in the most head-to-head matchups wins.
   Cedric        -- 51 -- First place
   Ava           -- 49
   Equal Support --  0
 Cedric wins.
   Runoff math:
     100  ballots cast
   −   0  Equal Support (no preference between the two finalists)
     ───
     100  voters with a preference  (majority = 51)
           Cedric 51 (51%)  ·  Ava 49 (49%)

──────────────────────────────────────────────────

[Bloc STAR: Round 3: Scoring Round]
 The two highest-scoring candidates advance to the next round.
   Ava           -- 294 -- First place
   Deegan        -- 227 -- Second place
   Eli           -- 224
 Ava and Deegan advance.

[Bloc STAR: Round 3: Automatic Runoff Round]
 The candidate preferred in the most head-to-head matchups wins.
   Deegan        -- 51 -- First place
   Ava           -- 49
   Equal Support --  0
 Deegan wins.
   Runoff math:
     100  ballots cast
   −   0  Equal Support (no preference between the two finalists)
     ───
     100  voters with a preference  (majority = 51)
           Deegan 51 (51%)  ·  Ava 49 (49%)

──────────────────────────────────────────────────

[Bloc STAR: Round 4: Scoring Round]
 The two highest-scoring candidates advance to the next round.
   Ava           -- 294 -- First place
   Eli           -- 224 -- Second place
 Ava and Eli advance.

[Bloc STAR: Round 4: Automatic Runoff Round]
 The candidate preferred in the most head-to-head matchups wins.
   Eli           -- 51 -- First place
   Ava           -- 49
   Equal Support --  0
 Eli wins.
   Runoff math:
     100  ballots cast
   −   0  Equal Support (no preference between the two finalists)
     ───
     100  voters with a preference  (majority = 51)
           Eli 51 (51%)  ·  Ava 49 (49%)

[Bloc STAR: Winners — Bloc STAR Voting Method (4 winners)]
 Bianca
 Cedric
 Deegan
 Eli

Then it happens again. And again. And again:

Seat	Scoring round	Runoff	Seat goes to
1	Ava 294, Bianca 231, Cedric 228, Deegan 227, Eli 224	Bianca 51 – Ava 49	Bianca
2	Ava 294, Cedric 228, Deegan 227, Eli 224	Cedric 51 – Ava 49	Cedric
3	Ava 294, Deegan 227, Eli 224	Deegan 51 – Ava 49	Deegan
4	Ava 294, Eli 224	Eli 51 – Ava 49	Eli

Winners: Bianca, Cedric, Deegan, Eli. Ava reaches all four runoffs as the top-scoring finalist and loses all four by the same two votes. BetterVoting agrees exactly — same four winners, same seat order, tieBreakType: "none" in all four rounds, no tie and no lot anywhere. The result is fully deterministic; there is nothing to appeal to.

Full count: bv1835_8h3yrx_score_leader_no_seat · two-view lesson: BV1835 · yaml

Why the 51–49 repeats¶

Because the two camps cancel and a tiny bloc decides. On Ava vs. anybody, the camps split 49–49: each camp scored Ava 3 and their own pair 5/4, so half the electorate prefers the other candidate and half prefers Ava. The 2-voter swing bloc scored Ava 0 and everyone else 5/4/3/2, so it breaks every one of those ties against her — 51–49, four times.

That also makes Ava the Condorcet loser: she loses every head-to-head matchup in the race. The engine says so in as many words, and adds the twist that she'd have won under Approval:

[Condorcet Loser]
  Condorcet Loser: Ava — loses every head-to-head matchup — elected by Approval!

[Divergence from STAR]
  STAR     = Bianca
  Approval = Ava   (differs from STAR)

So the point tally and the preference tally genuinely disagree here, and Bloc STAR follows the preference tally. Both times. Every time.

Is this a defect?¶

It depends on what you think the scoring round is for, and the honest answer is that both readings are coherent.

Reading 1 — working as designed. This is runoff reversal, STAR's signature step, repeated once per seat. The whole reason STAR appends a runoff to a score ballot is that raw point totals reward a candidate everybody tolerates over the candidate more voters actually prefer — and reward strategic minimization besides. The runoff asks the majority directly, and here the majority answered four times. A method that seated Ava would be Score voting, not STAR, and the repo's own Approval / Score critique is exactly the argument for not doing that.

Reading 2 — the multi-winner sting. In a single-winner race a runoff reversal costs the score leader the seat, and that is the deal you signed up for. Here it costs them all four, and there is no rung anywhere in Bloc STAR at which "led every round and lost every round" earns anything. A body that wanted its compromise figure in the room does not get her, no matter how many seats it has. If that is the goal, the fix is a different method — a Condorcet-flavored count like Bloc Ranked Robin, or a proportional one — not a complaint about the tabulation.

What the case is not is the majority sweep. Nobody sweeps here: the two camps take two seats each, a perfectly proportional-looking 2–2. The score leader alone is excluded, and she is excluded by the runoff, not by anybody's majority.

What to take away¶

Read the runoff line, not the score line. In a Bloc result the scoring round tells you who the finalists were; the runoff tells you who won. Reporting a Bloc race by its score totals will eventually publish the wrong winner.
A large score lead is not a safe seat. Ava's 63-point margin bought her four appearances in the runoff and nothing else. Margins in the two rounds are not commensurable and shouldn't be compared.
Being everyone's second choice is a weaker position in Bloc than it looks. The compromise candidate has to actually be preferred by more than half of the deciding voters in at least one pair — and against a field of factional favorites, they may never be.