Skip to content

Cross-method concepts — the ideas that belong to no single method

New here? Start with Ballot & Terminology Basics — a short four-step reading path through the ideas people most often get wrong (terminology, scores vs. ranks, what counts as "ranked," and strict vs. weak ranks).

Not every glossary term needs its own page — most are fine as a one-liner in Glossary — voting methods & criteria. This folder holds the handful of concepts that are load-bearing in debates and worth a focused page with a worked example and links to the test-case YAMLs that demonstrate them.

This folder is the cross-method half of the library. Ideas that belong to no single method live here — topics/ (flat concept pages plus thin per-topic hubs), the voting paradoxes, scores and ranks, the curriculum, the glossary, the tabulation engines, and the authoring canon (TIPS, templates).

Looking for one particular method? Its pages are not here — they live in that method's own folder, alongside its runnable examples, and the folder's README is its start-here:

Method Start here Concept pages
STAR 01_STAR 01_STAR/01_Learn/
Proportional STAR 03_STAR_PR 03_STAR_PR/01_Learn/
Approval 04_Approval 04_Approval/01_Learn/
Ranked Robin 05_Ranked_Robin 05_Ranked_Robin/01_Learn/
RCV-IRV 06_Other/RCV_IRV 06_Other/RCV_IRV/concepts/
Range / Score 06_Other/Range 06_Other/Range/concepts/

(Until 2026-07-29 those pages lived here, in a parallel 07_Concepts/<Method>/ tree — which meant every method had two front doors. Old links still work: every moved page keeps a permanent redirect.)

General & cross-method

Concept One line
Ballot & terminology basics the 4 ideas people most often get wrong — start here
Scores vs. ranks (don't confuse them!) order-only (ranks) vs. order+strength (scores) — relative vs. absolute preference
Scoring methods aren't RCV scoring methods (Approval, STAR) rate candidates and sit outside the ranked-voting family
Approval — honest limits no preference strength/order, the approval-threshold dilemma, bullet-voting; balanced with its equal-vote simplicity
Approval + Top-Two the St. Louis package — Approval primary, top-two general; why the runoff must be a second election (0/1 ballots re-counted head-to-head just echo the approval count), and how STAR folds it into one ballot
Strict vs. weak ranks many ranked methods allow equal ranks & compare pairwise — RCV-IRV does neither
"Preference" (a slippery word) everyday opinion vs. technical "ranking"; why "Preferential Voting" is a misnomer
Ranked Robin / RCV-RR (= Copeland) a Condorcet method: ranks + equal ranks, every pair compared head-to-head; tabulable via the pref_voting engine
Ranked Robin (RCV-RR) — honest limits Condorcet cycles need a resolution rule; no preference strength (bland-compromise winner); burial
Proportional Representation (STAR-PR methods · STV · the two, head to head) multi-seat — coalitions get proportional seats; the score family and STV land alike, Bloc differs (301.1 / 301.2)
Tabulation, step by step (201) the same ballots counted both ways — STAR's 2 steps vs IRV's elimination rounds
RCV-IRV vs. STAR (side-by-side) balanced comparison hub — real strengths on both sides — routing to the facet pages
Tabulation engines — BV, LH, RCV-IRV why an election has two reports (BetterVoting's visual display + the LH engine's text report), how they map, and the convert→validate→test pipeline; hub for the per-engine folders
STAR Reporting — reading & comparing results how a result is reported: scores, runoff, percentages, Equal Support/abstentions; LH vs BetterVoting and where they differ; the Score Distribution and Preference Matrix up close
The LH starvote engine what the engine is (a thin fork + a thick reporting wrapper) and our improvements — the minimal on-screen report vs the always-full _tabulated mirror, the matrix/divergence/runoff-funnel reporting, and multi-method dispatch
Reading a STAR report (201) the full engine report, section by section — matrix, divergence, both rounds, winner — and which parts to show 101 vs 201 vs 301
Quorum — did enough show up? a turnout threshold separate from who wins; abstentions count toward it; an unmet quorum means no winner — won the count but not elected

Method-specific concepts — in the method folders

The per-method concept indexes used to be duplicated here. They aren't any more: each method folder owns its own index, so there is one list per method instead of two that drift apart. Use the table at the top of this page, or go straight to the index you want:

Deeper conversation/debate scripts live beside their topics, indexed in Conversation scripts — index (e.g. Favorite Betrayal — Does Only RCV Avoid It?, "Exhausted Ballots" — What FairVote's Word Actually Hides); these pages are the shorter, reference-style explainers the glossary links to.