The Scoring Round¶
One line: STAR's first step. Add up every candidate's stars across all ballots; the two highest-scoring candidates become the Finalists and advance to the Automatic Runoff. That's the whole first round — simple addition. The scoring round measures how much support each candidate has; the runoff then decides how many voters prefer each finalist (Score Then Automatic Runoff).
→ Round 2: The Automatic Runoff · the whole report: Reading a STAR report · the two-step count vs IRV: Tabulation, step by step.
What the round does¶
- On each ballot, every candidate has a score from 0 to 5 (equal scores allowed; a blank counts as 0).
- Add up each candidate's scores across all ballots — plain summation.
- The two highest totals are the Finalists. Everyone else is done; only the top two advance to the runoff.
This is utilitarian aggregation — it rewards total support. A candidate rated a solid 4 by almost everyone can out-total a candidate given 5 by a passionate few and 0 by the rest. Whether that top-total candidate actually wins is the runoff's job, not the scoring round's.
Scoring Round
The two highest-scoring candidates advance to the next round.
Carmen -- 15 (average 5) -- First place
Andre -- 12 (average 4) -- Second place
David -- 8 (average 2+2/3)
Ella -- 1 (average 1/3)
Blake -- 0 (average 0)
Carmen and Andre advance.
Totals decide — averages are just for reading¶
The report prints each candidate's total and, in parentheses, the average per ballot. The total is what ranks them and picks the finalists; the average is a readability aid (it says "about how many stars per voter"). With every ballot scoring every candidate, total and average rank candidates the same way — but it's the total the engine advances on.
Equal scores are allowed — and honest scoring is safe¶
You may give two candidates the same score; you are never forced to invent a preference you don't feel (equal ranks / weak orders are a ranked-ballot problem STAR doesn't have). This is also what makes the scoring round resist vote-splitting: because you can give two similar candidates both a 5, running an ally no longer bleeds your favorite's total — the two don't split a shared pool of support the way they would under choose-one. → the spoiler effect. And you never need to exaggerate: in a pure score election the smart move is to give your favorite 5 and everyone else 0 (any in-between score just helps a rival's total), which collapses the ballot into Approval. STAR's runoff removes that pressure, so your honest 3s and 4s in the scoring round still do real work without costing your favorite. Why that matters, and its narrow limit: residual vote-splitting · STAR's honest limits.
Is the Scoring Round just Range (Score) voting?¶
Essentially yes — Round 1 is a Range / Score election. You score every candidate 0–5 and add up the totals, exactly as Score / Range voting does. The difference is what happens next:
- Range voting stops here and crowns the highest total. The scoring round is the whole election.
- STAR doesn't. The top two totals are only Finalists; the Automatic Runoff then elects the majority-preferred one.
That second step isn't a formality — it changes how you should vote. In pure Range, your ballot has the most leverage if you give only 5s and 0s (any middle score just helps a rival's total catch yours), so Range rewards min/max "bullet" voting and, if everyone does it, collapses toward Approval. STAR's runoff removes that reward: since the runoff only asks which finalist you scored higher, your honest 3s and 4s keep their meaning and you never have to exaggerate. So the scoring round is Range's expressive half; the runoff is what lets voters use that expressiveness honestly — STAR's hybrid nature.
Ties in the scoring round¶
Two kinds of tie can happen here:
- A tie for a Finalist slot (e.g. two candidates tie for 2nd). STAR resolves which one advances with the tie-break ladder — first the head-to-head pairwise result, then the five-star count, then, as a last resort, a documented lot.
- A rare three-way (or larger) tie for the top scores that the ladder can't split cleanly falls through to the lot.
Full ladder, lot order, and imported tieBreakType: STAR Tie-Breaking.
What it feeds — and what it can't decide alone¶
The scoring round only ever produces two finalists. It deliberately does not crown the highest total, because a big total can come from a passionate minority (why a second step at all?). The Automatic Runoff then asks the majority-protecting question the total can't — between these two, which do more voters prefer? When the score leader also wins the runoff (the usual case) the rounds agree; when they don't, that's a Runoff Reversal. Why STAR is built as this two-step hybrid: STAR's hybrid nature; the three different "winners" a report can name: three winner notions.
FAQ¶
"Does a blank count as 0, or as 'no opinion'?" As 0 — the lowest score. A blank is the same as scoring that candidate 0; there's no separate "skip this candidate" that removes them from the total.
"Why totals, not averages?" Since every ballot scores every candidate, total and average rank candidates identically — and the total is the honest measure of how much support across the whole electorate. The report prints the average only as a reading aid.
"Can I give two candidates the same score?" Yes — equal scores are allowed and never penalized. If both reach the runoff, your ballot is Equal Support between them.
"Should I give my favorite 5 and everyone else 0?" You can, but it usually forfeits your say in the runoff. If your favorite doesn't make the top two, all-0s means you've expressed no preference among the candidates who did — so give the "acceptable" candidates 3s and 4s so your full vote lands on the better finalist. Honest scoring is also the strong strategy. → Watch it backfire on real ballots: Bullet voting backfires shows the same nine voters twice, drawn ballot by ballot — honestly, and then with four of them zeroing their second choice. The zeroing works exactly as intended (it knocks the compromise out of the runoff) and then hands the election to the candidate those four scored 0.
"Does a candidate need a majority to advance?" No. The two highest totals advance, full stop — majority only matters in the runoff.
"Does scoring my second choice hurt my favorite?" In the scoring round, a little — and STAR accepts this by design. Giving a second choice a positive score adds to their total, which could in principle help them out-total your favorite. That means the scoring round is not Later-No-Harm compliant (RCV-IRV is; STAR trades it away). The upside is bigger: if your favorite doesn't reach the top two, that honest support keeps you a voice in the runoff, and scoring an ally doesn't let a rival split your side. (Note: Later-No-Harm is a different criterion from favorite-betrayal — STAR never rewards ranking someone above your true favorite.)
"Why not just crown the highest total and skip the runoff?" Because a big total can come from a passionate minority (lots of 5s) while most voters rated that candidate low. That's the whole reason for the second step — see Why a second step at all? and Runoff Reversal.
"Can precincts add up their own subtotals?" Yes — scoring-round totals are plain sums, so they're summable: precinct subtotals add to the statewide total, unlike RCV-IRV. → summability.
At a glance¶
| Scoring Round (Round 1) | Automatic Runoff (Round 2) | |
|---|---|---|
| Question | how much total support? | how many prefer each finalist? |
| Math | add every candidate's stars | one vote per ballot, to the higher-scored finalist |
| Output | the two finalists | the winner |
| Ties | tie for a slot → tie-break ladder | even split → tie-break ladder |
| Character | utilitarian (total support) | majoritarian (majority preference) |