Conversation scripts — index¶
A series of recorded Larry (host) ↔ Adam (expert) conversations that teach STAR Voting through Q&A plus live demos run from this repo. The scripts now live beside their topics — STAR episodes in STAR Voting — concept pages, RCV-IRV episodes in RCV_IRV/, and the foundational ones here in 07_Concepts/ — and this page indexes them. The demos come from 07_Concepts/ and the method folders' cases/ directories (mostly 01_STAR/ and method_comparisons/).
Start here: whats_so_good_about_STAR_Voting.md — the flagship episode that introduces the whole arc.
Format policy — scripts are for presenters; readers get educational pages (decided 2026-07)¶
The Larry↔Adam dialogue format exists for recording — it is a presenter artifact, not the reading surface. A visitor landing on a topic should meet a normal educational page, not a screenplay. So:
- Every topic's primary surface is an educational concept page. A conversation script survives only where the spoken format is the product (the flagship pitch, the personal cold open) — and even then it should point to the educational pages for the substance.
- The model pattern is what_is_a_voting_method.md: a concept page (voting_method_ballot_and_count.md) is primary, and the script opens by saying it is the spoken/video companion.
- Status of the dialogue-format pages:
| Page | Verdict | State |
|---|---|---|
| are_equal_score_votes_discounted.md | convert | ✔ converted to an educational page (2026-07) |
| what_is_a_voting_method.md | keep as companion | ✔ has its concept-page twin (the model) |
| whats_so_good_about_STAR_Voting.md | keep as script | the flagship recording script; substance lives in STAR — start here & benefits |
| ~~why_do_you_love_STAR_Voting.md~~ | retired 2026-07-29 | a first-person cold open that put personal statements in Adam's mouth. Removed at his request; the site URL redirects to the flagship script above |
| our_voting_system_is_broken.md | convert | ✔ done (2026-07) — converted to an educational "problem with plurality" page (dialogue → prose; demos as links) |
| favorite_betrayal_voting_301.md | convert | ✔ done (2026-07) — educational reference body; §7 recast from a Larry↔Adam recording script into an RCV-advocate ↔ STAR-advocate debate (debate-prep, not a presenter artifact) |
| RCV_or_IRV_whats_the_right_word.md | convert | ✔ done (2026-07) — now a short concept page (dialogue → prose) that defers the full taxonomy to canonical TIPS_terminology |
| exhausted_ballots_301.md | convert | ✔ done (2026-07) — reference body kept; §7 recast from a Larry↔Adam script into an RCV-advocate ↔ STAR-advocate debate (matches favorite_betrayal_301) |
Where to start & how to organize¶
Record one flagship episode first (the "What's so good…" script), because it forces you to nail the core arc end-to-end. Then break that arc into short, single-topic clips you can release on their own and reuse. The flagship doubles as the table of contents.
Format that works for recording: Larry asks the naive-but-smart question; Adam answers in 2–4 sentences; then a live demo or a slide. Keep Adam's turns short — it's spoken, not a lecture. Every segment should end on either a demo or a memorable one-liner.
Proposed episode roadmap¶
| # | Episode (clip) | Core question | Repo demo | Source slides |
|---|---|---|---|---|
| 00 | What is a voting method? (the foundation) | ballot vs count; plurality vs majority | split_voting/00_plurality_vs_majority |
"Rating vs Ranking", "Plurality vs Majority" |
| 0 | Why do you love STAR? (cold open) | the personal "why" | — | Full Deck title + dilemma |
| 1 | What's so good about STAR? | the whole pitch in one arc | flagship file's demos | Full Deck (all) |
| 1.5 | Is our voting system broken? (Problem with Plurality) | is it broken — frequent? serious? worsening? | split_voting/01*, 04* |
"Problem: Vote-Splitting", "The Voting Dilemma" |
| 2 | The spoiler effect | why Choose-One breaks with 3+ | split_voting/01*, 03* |
"Vote-Splitting", "Domino Effect" |
| 3 | Ranks vs Scores | why scores carry more info | (whiteboard) + 00a |
"Rating vs Ranking" |
| 4 | The STAR ballot | how do I vote? | — | "Instructions" / 5-star ballot |
| 5 | The two rounds | why Score then Runoff? | 00a_c2_b1, vote_splitting2 |
"How STAR Works", Dual-Nature doc |
| 6 | Benefits of STAR | Simple/Honest/Expressive/Accurate/Equal | accuracy demos | "What do we want", accuracy chart |
| 7 | STAR vs RCV-IRV | pros & cons, honestly | 4-cand [Divergence from STAR] demo |
Scorecard, Burlington, Alaska |
| 7.5 | RCV or IRV? (nomenclature, 201) | which word, and why | — | "About Ranked Choice" |
| 8 | Counting & trust | summability, audits | any file's matrix | "Summability" slides |
| 9 | Multi-winner & proportional | electing several seats | 02_Multi_winner/* |
"Proportional Representation" |
| 10 | The honest limits | "resistant, not proof" | — | "Voting Theory Facts" |
| 11 | "Discounted votes?" (objection) | are equal-score votes ignored? | page · equal_support_runoff_demo.yaml |
"No-Preference Votes in the Runoff" |
| 12 | Favorite betrayal (Voting 301) | does only RCV avoid it? | — | "RCV Common False Claims", "Alaska '22", pass/fail criteria |
| 13 | "Exhausted ballots," untangled (301) | what FairVote's word actually hides | equal_support_runoff_demo (contrast) |
"Ranked Choice Deal Breakers", "Alaska '22", Exhausted Ballots deck |
| ★ | RCV-IRV false-claims index (reference, not a clip) | look up any "RCV" claim → its precise rebuttal | — | "RCV Common False Claims" |
Episodes 1, 2, 5, and 7 are the highest-value if you only make four. Episode 11 (are_equal_score_votes_discounted.md, now an educational page — build the clip's Q&A from its sections) is debate-prep — deploy it only when a skeptic raises the FairVote "discounted votes" line; don't lead with it.
Suggested additional topics (good standalone clips)¶
- "Equal Preference / no-preference in the runoff" — what an equal score does.
- "Bullet voting — is it a strategy?" (it isn't; show why).
- "The preference matrix, read out loud" — using the matrix the engine prints by default.
- "Abstentions, blanks, and 0s" — what the ballot markers mean (use the engine's legend + the
Absscore-distribution bucket).
Cross-referencing the presentations¶
You have four decks/docs in play. Treat them as source slides, not as the script — pull individual slides into the relevant segment.
| Short name | What it's best for | Link |
|---|---|---|
| Full Deck 2025 | the canonical slide library; nearly every segment | docs.google.com/presentation/d/1wyR_1aSlATkOzcbGhWCD0_mRtRK60PDS3INIazhQl1w |
| Automatic Runoff | deep-dive visuals for Segment 5 (the runoff) | docs.google.com/presentation/d/1DNJg5kbsNbNQFV3bosm4KZmgJtNfIbQdpirhm5XlCqY |
| Torrance LWV (Arend) | the cleanest single-talk flow + the scorecard table | docs.google.com/presentation/d/1xPVPWnqJJPrkdm68uuZ9dYKqCsPqV1zgOcRnRBhZaAM |
| Dual Nature (doc) | the "strength vs numbers" framing for Segment 5 | docs.google.com/document/d/1n3pkr_S9MBucJMg7ANhk5ZZ00OLTW3Kj5Bp-vruW10k |
Turning Arend's Torrance deck into Q&A: it already moves in a teachable order (problem → spoiler → alternatives → RCV facts → STAR → scorecard). The best approach is to keep his slide order as the spine and convert each slide title into Larry's question: - "PROBLEM: VOTE-SPLITTING" → Larry: "What actually goes wrong with more than two candidates?" - "ABOUT RANKED CHOICE" → "People love RCV — what does it get right, and where does it fall down?" - "Single-Winner Voting Method Scorecard" → "Can you put STAR, RCV, and Choose-One side by side?" (then show the table as the answer). Each slide becomes one Q→A→[show slide] beat. His scorecard slide is the single strongest asset you have for Segment 7 — lead the comparison with it.
For the "Dual Nature" doc you need to turn into a presentation: it's already structured as an argument (cardinal phase / ordinal phase / why it matters / strategy). Convert each bold heading into one slide and one Larry question: - "Scoring Phase (Cardinal)" → slide: strength of support. - "Runoff Phase (Ordinal)" → slide: number of supporters. - "Why This Matters" → slide: cardinal is expressive but gameable; ordinal resists gaming but splits votes; STAR uses each to cover the other's weakness. - Close on the Tim Peters quote — it's a great mic-drop slide. This is Segment 5's deck; build it once and reuse it.
Suggested changes / new presentations¶
- New: a 5-slide "Scores vs Ranks" micro-deck (Segment 3). None of the existing decks isolate this cleanly, and it's the conceptual hinge.
- New: a "Live Demo" deck that is mostly blank frames cueing you to switch to the STAR engine — so the recording cuts cleanly between slide and terminal.
- Change to the Full Deck: it has several near-duplicate "vote-splitting" slides with identical speaker notes. For the recordings, make a trimmed recording cut that keeps one of each, so you're not paging past repeats.
- Add a Burlington + Alaska "real failures" two-slide pair to the front of Segment 7 — concrete beats abstract.
Keeping presentations & Google Doc links in sync¶
Google Slides/Docs don't auto-sync with this repo, so don't try to mirror their content here. Instead keep a single source-of-truth link registry and a mapping, and let each live in its own home:
- One
LINKS.md(I can create it) listing every deck/doc: short name, URL, one-line purpose, and "last reviewed" date. Every script links by short name, never by raw URL — so when a URL changes you fix it in one place. - Map, don't copy. Scripts reference slides as
[SLIDE] Full Deck — "title". That survives slide reordering far better than slide-number links. - Stable anchors. In Slides, give key slides a short, searchable title (you already do). Searching the title is more durable than a
#slide=id…deep link, which breaks on copy/duplicate. - One-way canonical direction. Pick the canonical artifact per topic (e.g. Arend's deck is canonical for the scorecard) and have others point to it rather than duplicating it.
- Review cadence. When you edit a deck, bump its "last reviewed" date in
LINKS.mdand grep the scripts for its short name to catch stale references.
The repo's own files (engine output, example YAMLs) are canonical and version- controlled — so prefer them for anything you want to stay exactly in sync (the live demos), and treat the Google decks as the visual layer on top.
Demo software: BetterVoting vs the STAR engine¶
(From your "pros and cons" note — summarized with a recommendation.)
Use both, in this order:
- Drive the teaching with the STAR engine in this repo (the
starvote_…fork). It's dynamic: you can change a ballot on the fly to make a scenario land, the reporting options show exactly what you want (Condorcet, RCV-IRV, Approval, the vote-splitting check), and every example is a saved, repeatable file. That's perfect for "watch what happens when I move this one score." - Send the audience to BetterVoting.com to participate. Beautiful graphics, a real shareable ballot link, and audience engagement (have them vote on their phones). That's the call-to-action moment.
Why not BetterVoting for the teaching demos: casting each ballot one-by-one is slow and error-prone on camera, you can't tweak scores live, the sandbox's reporting is incomplete, and (per your notes) it mishandles ties — which is exactly the kind of edge case a teaching demo wants to be correct about.
Make the engine demos better for recording:
- Pre-stage the exact files (see the flagship's demo index) so you never type ballots live.
- The engine's default on-screen report already shows the matrix, the runoff summary, and the [Divergence from STAR] block with the rest quiet; when you'll narrate a heavier section (Condorcet lines, the score distribution), run with --full — so the screen isn't busy.
- Keep a blank "Live Demo" slide to cut to, so the video editor has a clean seam.
- For the spoiler beat, the [Vote-splitting check] block gives you the punchline in numbers — read it out loud.
Honest caveat from your own note: after the engine demo you'll still point people to BetterVoting to actually try it — so frame the engine as your teaching instrument and BetterVoting as their hands-on tool. They're complementary, not competing.
Files in this folder¶
whats_so_good_about_STAR_Voting.md— the flagship Larry↔Adam script.README.md— this plan (roadmap, slide map, sync strategy, demo software).