The Automatic Runoff — slide outline¶

A presentation deck for STAR's second step, built to scale 101 → 201 → 301: open plain-language for a general audience, then add the machinery (reversal, Equal Support, two denominators) for a technical/debate crowd. Stop wherever your audience is.

Format: one slide per ## heading; bullets are the on-slide text, > Notes: is the speaker note (don't put it on the slide).
Live results: every example slide links the real BetterVoting result so you can show it on screen, plus the matching repo lesson (BV screenshots beside the LH text report).
Concept hub backing this deck: The Automatic Runoff Round · the worked teaching set: Runoff Reversal · presenter's guide: Teaching Runoff Reversal — a step-by-step guide.

1 · Title — The Automatic Runoff¶

STAR = Score Then Automatic Runoff. The runoff is the "AR."
One promise to plant up front: you can score honestly — and the majority still wins.
Today: how much vs how many, why the top scorer can lose, and what that buys you.

Notes: Say the acronym out loud and unpack it — most people hear "STAR" and don't know the runoff is half of it. Keep the room's eyes on one phrase: how much vs how many.

2 · The ballot (60 seconds)¶

Score every candidate 0–5 stars — like rating movies. High = love, 0 = no.
You can give the same score to more than one (no forced ranking).
That's the whole voter experience. Everything next is just counting.

Notes: For a public audience this is the only "how to vote" slide. The runoff is a counting rule — voters never do anything extra for it.

3 · STAR counts in two steps¶

Step 1 — Scoring Round: add every candidate's stars. Top two = Finalists.
Step 2 — Automatic Runoff: look only at those two. Each ballot's full vote goes to whichever finalist it scored higher. The finalist more voters prefer wins.
Two steps, hand-countable. That second step is the whole point of the method.

Notes: Resist explaining why yet — just establish the mechanism. The "why" is the next slide. Concept: The Automatic Runoff Round.

4 · Why a second step at all? — how much vs how many¶

Scores measure how much support a candidate has (points).
The runoff counts how many voters prefer each finalist (people).
Points and people aren't the same thing — a few very high scores can inflate a total.
The runoff makes the winner the majority's choice, not the biggest pile of stars.

Notes: This is the spine of the whole talk. Everything else is an illustration of how much vs how many. If they remember one phrase, make it this one.

5 · The runoff keeps the ballot honest¶

Without a runoff, the smart move is to score only your favorite 5 and everyone else 0 — any in-between score just helps a rival.
If everyone does that, the 0–5 ballot collapses to 0s and 5s — that's Approval voting.
STAR's runoff removes that incentive: it only asks which finalist you scored higher, so your honest 3s and 4s still do real work.
The runoff is what lets STAR keep an expressive ballot people can use honestly.

Notes: The "strategic scoring degenerates to Approval" argument. Powerful with a technical crowd; for a public crowd, the one-liner is "you never have to exaggerate." Glossary: Approval.

6 · What the runoff does — three buckets¶

For each ballot, compare the two finalists' scores on that ballot:
scored one higher → a full vote For that finalist;
scored them the same → Equal Support (no preference between these two).
Finalist with more votes wins. Losers of the score round get no second count.
The result names all three buckets, then states the decisive split.

Notes: "Equal Support" gets a full slide later (slide 13). Flag it here, don't dwell. Concept: What the round does.

7 · The flow (one picture)¶

Scoring Round → Top two → each ballot picks its preferred finalist → tally → winner.
Branch the audience can read off it: if the winner isn't the score leader, that's a Runoff Reversal — the safeguard firing, not a glitch.
Equal Support leaves the percentage but still counts in the scores.

Notes: Show the mermaid flowchart from the concept hub. Don't narrate every node — point at the one fork that matters (winner = leader? yes/no). Diagram: The flow.

8 · Live baseline — the runoff confirms the leader¶

Runoff 01 (control). Aspen leads the Scoring Round (12) and wins the runoff 2–1.
Most of the time the score leader is also the majority's pick — the runoff just checks.
Show this first so the next slide's reversal can't look "rigged."

Notes: Open the BV link live; point at the dashed majority line Aspen clears. Live result: https://bettervoting.com/r2pvc9/results Lesson: Runoff 01 — confirms the leader.

9 · The key move — the top scorer can lose¶

A score total can't tell your #1 from your strong #2 — a 4 looks the same either way.
So a candidate who's everyone's solid second choice can pile up the most stars without being anyone's first choice.
The runoff forces the question the total can't answer: between these two, which is #1?
When how much and how many point at different candidates → Runoff Reversal.

Notes: This is the hardest idea for learners. The "a 4 can't tell #1 from a strong #2" line does most of the work — say it slowly.

10 · Live reversal — the atom¶

Runoff 02 (smallest reversal). Everyone likes Austin (most stars, 13) — but two of three voters prefer Boston, so Boston wins the runoff 2–1 (67/33).
Same machinery as the baseline; only the ballots changed.
BetterVoting itself pops up "Why is the top-scoring candidate different from the winner?"

Notes: Put slide 8 (Aspen) and this side by side if you can — identical mechanism, opposite outcome, because the preferences differ. Live result: https://bettervoting.com/yx9447/results Lesson: Runoff 02 — the atom.

11 · It's not a small-numbers fluke — at scale¶

Runoff 04. The atom blown up to nine voters: Maple leads on stars (39), but Olive wins the runoff 6–3 — a clean 2-to-1 majority (67/33).
Adding voters in the same proportions doesn't rescue the score leader — it makes the majority's verdict louder.
How many is a headcount; headcounts grow with the crowd, a few enthusiasts' extra stars don't.

Notes: Use this to kill the "cute toy example" objection. Same shape, bigger room. Live result: https://bettervoting.com/bfjqmg/results Lesson: Runoff 04 — reversal at scale.

12 · The other face — intense minority vs majority (201)¶

Runoff 03. Two voters love Dakota (5s) → Dakota tops the scores (22). But three others prefer Eden, who wins the runoff 3–2.
A passionate minority can top the score chart; it can't outvote the majority in the runoff.
Where Runoff 02 had a broadly-liked leader, this is an intense-minority leader — same reversal, opposite flavor.

Notes: First 201 slide. Good place to say: STAR rewards breadth of support, not just intensity. Eden is also the Condorcet winner — a calm case, no cross-method drama. Live result: https://bettervoting.com/rkgtpk/results Lesson: Runoff 03 — enthusiasts vs majority.

13 · Equal Support — a cast vote with no preference (201)¶

Scored the two finalists the same? You have no preference between these two — that's Equal Support, set aside from the runoff %.
It is not an abstention: it counts fully in the Scoring Round; only a fully-blank ballot abstains.
Runoff 05. Rosa leads (21), but two voters rate Rosa = Sage, so 3 of 5 decide — and they pick Sage 2–1.

Notes: The single most-confused term in STAR. Hammer "cast vote, no preference ≠ didn't vote." This is the bridge to the two-denominator slide. Live result: https://bettervoting.com/xgkw3w/results Lesson: Runoff 05 — reversal with Equal Support · term: GLOSSARY.

14 · Two denominators — how many of whom (201)¶

The winner needs a majority of voters who expressed a preference between the two finalists — not a majority of everyone.
% Between Finalists = out of the decided voters; % Runoff Votes = out of all ballots. Same votes, two denominators.
LH names its denominator inline: "Voters with a preference: 3 of 5 (2 Equal Support)."

Notes: Show the BV Runoff Table (both % columns) and the LH one-line summary together — they're the same numbers, different words. Detail: Runoff percentages — two denominators · why the words differ: reporting_diff_BV_LH.

15 · A second control — the leader confirmed at scale¶

Runoff 06. Wren leads the scores (21) and wins the runoff 4–1.
Bookend to the reversals: the runoff doesn't change the answer here — but it asked the question, which is the safeguard.
A reversal only happens when how much and how many point at different candidates.

Notes: Reassures the room the runoff isn't out to overturn leaders for sport. It confirms far more often than it reverses. Live result: https://bettervoting.com/d664xw/results Lesson: Runoff 06 — confirms at scale.

16 · Variations to know¶

Exact tie in the runoff → even split → tie-break ladder (Scoring Round, then a documented lot). On a true 50/50 the % line is suppressed.
Two candidates → the Scoring Round is a formality; the runoff is the election.
Everyone gives everyone 5 → all Equal Support → tie → ladder decides (the silly case).

Notes: Keep this brisk — it's the "edge cases exist and are handled" slide. Detail: Variations.

17 · Devil's advocate — the questions you'll get¶

"Most stars = most popular, how can he lose?" — Most stars ≠ most preferred. Stars say how much; the runoff counts how many.
"Then just crown the highest total!" — Then everyone votes 5s-and-0s and STAR becomes Approval. The runoff keeps scoring honest and guarantees a majority winner.
"Isn't it unfair to the score leader?" — They didn't win; they earned a spot in the runoff. Leading the scores makes you a finalist, not the office-holder.
"Could the winner be almost nobody's favorite?" — No: preferred head-to-head by a majority of decided voters.

Notes: These are verbatim the questions audiences ask. Full list with answers: Quick questions.

18 · What the runoff buys you¶

A majority-backed winner — beats the runner-up among voters with a preference.
Honest scoring is safe — no reward for exaggerating; the 0–5 ballot doesn't degrade to Approval.
Resists intensity capture and center squeeze — a loud minority can't win on volume; a broadly-acceptable middle candidate isn't eliminated early (unlike IRV).
Still simple — two steps, easy to hand-count.

Notes: "Center squeeze" is an IRV-specific failure — name IRV precisely here, not "RCV." STAR's runoff is what avoids it. Comparison: center squeeze, STAR vs IRV.

19 · …and its limits (301)¶

It only considers the top two: a candidate who'd beat both finalists but finished 3rd never enters — so STAR can, rarely, miss the Condorcet winner.
The reversal surprises people; unexplained, it can feel illegitimate (the cost these lessons exist to fix).
One extra step + the Equal-Support wrinkle (two denominators) over a plain score count.
Like every method, it can't satisfy every fairness criterion at once (Arrow/Gibbard).

Notes: Don't hide the limits — naming them honestly builds credibility. Keep the Condorcet-miss for 301; it's not a beginner's first lesson. Detail: three winner notions.

20 · Reading the real reports — BV vs LH (201/301)¶

BetterVoting and the LH engine show the same numbers in different words — match them once and either report is readable.
LH folds BV's Runoff Votes + % Between Finalists + Equal Support into one self-reconciling line that names its denominator.
Every example in this deck has both views side by side in its lesson file.

Notes: For an audience that will actually run elections. Skip for a pure intro talk. Detail: reporting_diff_BV_LH · the full set: Runoff Reversal.

21 · A live edge case — when a tool gets it wrong (301, WIP)¶

Runoff 07. A flat 3,3,3 ballot (every candidate equal) is Equal Support — a cast vote. BetterVoting currently mis-files it as an abstention, dropping it.
Same winner (Blair), but BV's counts and score totals don't reconcile with a hand count of all four ballots.
Tracked as bettervoting#1407; kept as a teaching illustration until fixed.

Notes: Optional 301 capstone — shows the audience how to audit a result and why the Equal-Support vs abstention distinction (slide 13) actually matters in practice. Live result: https://bettervoting.com/tf73v9/results Lesson: Runoff 07 — flat ballot / BV bug (WIP) · the bug in full: When "no preference" gets called an "abstention".

22 · Recap — one slide to remember¶

STAR = Score Then Automatic Runoff. The runoff is the second half.
Scores = how much. Runoff = how many. The majority's pick wins.
The runoff lets you score honestly and still get a majority winner — and when it overturns the leader, that's the safeguard working.

Notes: End on the promise you opened with (slide 1). If you only had three sentences, these are them. Concept hub: The Automatic Runoff Round · teaching guide: Teaching Runoff Reversal — a step-by-step guide.

Appendix — every example at a glance¶

#	Slide	Level	BetterVoting result	Repo lesson
01	confirms the leader (control)	101	r2pvc9	Runoff_01
02	the atom (smallest reversal)	101	yx9447	Runoff_02
03	enthusiasts vs majority	201	rkgtpk	Runoff_03
04	reversal at scale (67/33)	101	bfjqmg	Runoff_04
05	reversal with Equal Support	201	xgkw3w	Runoff_05
06	confirms at scale (control)	101	d664xw	Runoff_06
07 ⚠️	flat ballot / BV bug (WIP)	301	tf73v9	Runoff_07

Suggested cuts by audience: public intro (101) → slides 1–11, 15–18, 22. Technical / debate (201+) → all of it. Lightning (8–10 min) → 1, 3, 4, 5, 9, 10, 18, 22.