Same opinions, every method — the line-up¶
The clearest way to see what a voting method is for: hold the voters' honest opinions fixed, change only the counting rule, and watch where each one lands. Nobody changes their mind; the only variable is how much of the ballot the method reads. Line the results up, and the virtues show themselves.
→ Its ballot-side twin — the same voter on three ballot styles: one voter, three ballots. The core distinction underneath both: scores vs. ranks.
One electorate, honest opinions¶
Five coworkers pick the team lunch — Sushi, Tacos, or Pizza. Everyone scores each option 0–5, honestly (this is the repo's canonical lunch vote, runnable · live on BetterVoting ↗):
| Voter | Sushi | Tacos | Pizza |
|---|---|---|---|
| Sofia | 5 | 0 | 3 |
| Sam | 5 | 0 | 3 |
| Tara | 0 | 5 | 3 |
| Theo | 0 | 5 | 3 |
| Pat | 3 | 1 | 5 |
Read it and the shape is plain: two love Sushi, two love Tacos, and everybody is happy with Pizza — nobody scores it below 3. Pizza is the option the whole table can live with; Sushi and Tacos each split the room in half.
The same opinions, as each ballot¶
Those honest scores contain every other ballot inside them. The voters didn't change — only what the ballot lets them say:
| Voter | Choose-One (top pick) | Ranking (their order) | Approval (all they're OK with) | Score |
|---|---|---|---|---|
| Sofia | Sushi | Sushi › Pizza › Tacos | Sushi, Pizza | 5·0·3 |
| Sam | Sushi | Sushi › Pizza › Tacos | Sushi, Pizza | 5·0·3 |
| Tara | Tacos | Tacos › Pizza › Sushi | Tacos, Pizza | 0·5·3 |
| Theo | Tacos | Tacos › Pizza › Sushi | Tacos, Pizza | 0·5·3 |
| Pat | Pizza | Pizza › Sushi › Tacos | Sushi, Tacos, Pizza | 3·1·5 |
First choices alone: Sushi 2, Tacos 2, Pizza 1. Notice Pizza looks weakest by first choices — precisely because it's everyone's happy second, not their passion.
The line-up — who wins under each method¶
Now count the same five ballots every way. (Every winner here is engine-verified on the runnable file above.)
| Method | What it reads | Winner |
|---|---|---|
| Choose-One (Plurality) | first choices only | Sushi |
| RCV-IRV (Hare) | first choices, then eliminate the lowest | Sushi |
| Approval | how many are OK with each | 🍕 Pizza |
| Ranked Robin | every head-to-head matchup | 🍕 Pizza |
| STAR | total scores, then an automatic runoff | 🍕 Pizza |
What the line-up shows¶
Three different whole-ballot methods — Approval, Ranked Robin, STAR — independently land on Pizza, the choice the whole table is happy with. The two methods that look only at first choices pick Sushi, which two of the five coworkers rated a flat 0.
Nobody voted strategically. Nobody changed their opinion. The entire difference is how much of the ballot the method reads:
- Choose-One and IRV decide on first choices, so the broadly-liked compromise — nobody's #1 — never gets its due, and the winner is an option half the room rejected.
- Approval, Ranked Robin, and STAR read the whole ballot — approval thresholds, every head-to-head, or scores plus a runoff — and each finds the candidate with real, broad support.
That's the case for expressive ballots and whole-ballot counts, in one lunch order: give voters room to say what they think, then count all of it, and the winner is the one a majority is genuinely glad about. STAR and Ranked Robin are two roads to that same destination.
Use this view anywhere¶
The line-up is a reusable lens, not a one-off. Any election in this library can be run through it — the point is always the same: hold the opinions fixed, vary the count, and let the method reveal itself. Bigger, real-world line-ups:
- Alaska 2022 — one electorate, four counts — a real federal race where the methods split (STAR & Ranked Robin find the consensus candidate; the first-choice methods don't), and its general-election counterpart where they all agree.
- Burlington 2009 — the same lens on real ranked ballots.