Skip to content

How the LH engine reports a STAR result

One line: the LH starvote engine prints a full text audit report — every number that produced the winner, in order — and saves a maximal _tabulated.txt copy of it. This page is the reporting-lens summary; the section-by-section walk is Reading a STAR report.

→ Hub: STAR Reporting · the engine itself: BetterVoting and the LH engine · GLOSSARY.

What the report shows, top to bottom

  • Tabulating N ballots. — the count, plus a note for any abstention (a fully blank ballot; an explicit all-zero is a cast ballot, not an abstention).
  • Score Distribution (optional) — per-candidate spread of 0–5 scores, with an Abs column for blanks.
  • Scoring Round — total stars per candidate; the top two advance.
  • Preference Matrix — head-to-head table, legend For – Equal Support – Against, and the Condorcet check.
  • Automatic Runoff Round — each finalist's vote count plus Equal Support (the ballots that scored the two finalists equally).
  • The runoff percentage line — on screen, a self-reconciling two-line summary:
Voters with a preference: 363 of 461 (98 Equal Support).
Dog 190 (52%) vs Cat 173 (48%); majority = 182.

It names the decided count against the total with the Equal Support gap inline, so the denominator never has to be inferred. (→ the two-denominator idea: Runoff percentages.) - "Runoff math" funnel — in the saved _tabulated.txt, the same line expands so the arithmetic is explicit:

   Runoff math:
     461  ballots cast
   −  98  Equal Support (no preference between the two finalists)
     ───
     363  voters with a preference  (majority = 182)
           Dog 190 (52%)  ·  Cat 173 (48%)
- Condorcet check / [Divergence from STAR] — whether the pairwise winner matches, and a flag when methods would disagree. - Winner.

On-screen vs the saved copy

The on-screen report is the engine's built-in default render — the defaults are the house style, so case files carry no options: block (a file may still set one to override any default). The _tabulated.txt sibling is always the maximal render — every analysis on, the runoff line expanded into the funnel — so the saved audit copy is fully self-explanatory; the --full CLI flag puts that same everything-on render on screen. Which flag shows which section: LH reporting options.

See it on a real election: the pet race report.