Skip to content

Alaska 2022 — one real electorate, four counts, two winners

▶ Live on BetterVoting: vote · results ↗ (election k3fmwv, Test ID BV2213 — all four races).

▶ Runnable companion to a published analysis. This case makes Clelland (2023), Ranked Choice Voting and Condorcet Failure in the Alaska 2022 Special Election hands-on. She worked out — by hand, from the full Cast Vote Record — what STAR and Approval would have done in this election. Here you can click the tabulated results and run the ballots yourself, and check her conclusions against a live engine. (Her ballot profile, our reduced model, and Graham-Squire & McCune all agree.)

Which Alaska election is this? The August 2022 US House special — the one where RCV-IRV misfired. It is not the November general (which worked fine — that's the runnable counterpart). Confused by all the Alaska elections? See the map: Alaska's RCV elections — the whole picture.

One line: in the August 2022 Alaska US House special election, Begich beat both other candidates head-to-head — a majority preferred him to Peltola and to Palin — yet RCV-IRV eliminated him first for having the fewest first choices, and Peltola won. Count the same ballots with a method that reads the whole ballot — Ranked Robin or STAR — and Begich, the Condorcet winner, wins. That is the center squeeze, in a real federal race.

This folder reproduces it as a runnable, reduced 200-voter model so you can tabulate it yourself and see all four methods diverge.

Learn it in three levels

New to this, or the "how did the guy who beat everyone lose?" part isn't clicking yet? Work up through these — each ends with "challenge your understanding" questions so you can check you actually get it:

  1. 101 — what happened, in plain language (zero jargon; the core intuition)
  2. 201 — the vote math, round by round (the exact IRV transfers, exhaustion, and head-to-head grid)
  3. 301 — every pathology, and is this fair? (spoiler / monotonicity / exhaustion — and the honest question of whether it's fair to STAR)

Provenance — a faithful scaling of the real profile

This is not the raw Alaska ballots; it is a ~943:1 reduction of the official preference profile (Table 1 of Graham-Squire & McCune, An Examination of Ranked Choice Voting in the United States, 2004-2022, arXiv:2301.12075). Every one of the nine real ballot types is preserved:

Real ballots (paper Table 1) Reduced Scores (Peltola, Begich, Palin)
47,407 Peltola › Begich › Palin 50 5,4,0
23,733 Peltola (bullet) 25 5,0,0
4,647 Peltola › Palin › Begich 5 5,0,4
34,078 Palin › Begich › Peltola 36 0,4,5
21,237 Palin (bullet) 23 0,0,5
3,659 Palin › Peltola › Begich 4 4,3,5
27,070 Begich › Palin › Peltola 29 0,5,4
15,478 Begich › Peltola › Palin 16 4,5,0
11,262 Begich (bullet) 12 0,5,0

Scores are a rank→score modeling overlay (favorite 5, then 4/3, unranked 0). The reduced model reproduces every headline figure within rounding. Companion visualizer: Equal Vote's Real RCV.

The four counts

Method Winner Why
Choose-One (Plurality) Peltola Most first choices: Peltola 80, Palin 63, Begich 57
RCV-IRV (Hare) Peltola Begich (fewest firsts) eliminated; 12 ballots exhaust; Peltola beats Palin 96–92
Ranked Robin (Condorcet) Begich Beats Peltola 93–84 and Palin 107–68 — wins every matchup
STAR Begich Score finalists Begich (641) & Peltola (480); Begich wins the runoff 93–84

Two of the four elect Peltola; the two that read the whole ballot elect Begich, the candidate a majority actually preferred. IRV is the lone method that fails the Condorcet winner — the center-squeeze signature. BetterVoting and an independent open-source count agree on all four winners, every race deterministic (tieBreakType: none).

View 1 — on BetterVoting (the live election)

Four races on the one electorate, live: bettervoting.com/k3fmwv/results. BetterVoting's own counts return Plurality Peltola · IRV Peltola · Ranked Robin Begich · STAR Begich, matching the independent count below exactly. Frozen export: bv2213_k3fmwv_alaska_2022_bv_export.json.

View 2 — an independent count (the full tabulation)

The same ballots, re-counted by the open-source engine this repo runs — Larry Hastings' starvote, extended here (added reporting and the RCV-IRV / Ranked Robin comparisons) — so you can verify the result yourself instead of taking the platform's word for it. It's tagged "LH" (Larry Hastings) elsewhere in this repo.

The score model tabulates STAR and auto-derives the other three counts (the [Divergence from STAR] block):

[Divergence from STAR]
  STAR                   = Begich
  Choose-One (Plurality) = Peltola   (differs from STAR)
  RCV-IRV                = Peltola   (differs from STAR)
  Note: no ballots had tied scores, so RCV-IRV vs STAR here is a genuine
        method difference, not a tie-breaking artifact.
  Note: Ranked Robin (RCV-RR) agrees with STAR, so RCV-IRV is the lone
        outlier — the classic center-squeeze signature.
  Full round-by-round reports (generated for review):
  RCV-IRV rounds: cases_tabulated/bv2213_k3fmwv_alaska_2022_RCV-IRV_tabulated.txt

--- STAR Voting Method (single winner) ---

[STAR Voting]
 Tabulating 200 ballots.
Count × Peltola,Begich,Palin
   50 ×       5,     4,    0
   36 ×       0,     4,    5
   29 ×       0,     5,    4
   25 ×       5,     0,    0
   23 ×       0,     0,    5
   16 ×       4,     5,    0
   12 ×       0,     5,    0
    5 ×       5,     0,    4
    4 ×       4,     3,    5

[STAR Voting: Scoring Round]
 The two highest-scoring candidates advance to the next round.
   Begich        -- 641 -- First place
   Peltola       -- 480 -- Second place
   Palin         -- 451
 Begich and Peltola advance.

[STAR Voting: Automatic Runoff Round]
 The candidate preferred in the most head-to-head matchups wins.
   Begich        -- 93 -- First place
   Peltola       -- 84
   Equal Support -- 23
 Begich wins.
   Runoff math:
     200  ballots cast
   −  23  Equal Support (no preference between the two finalists)
     ───
     177  voters with a preference  (majority = 89)
           Begich 93 (53%)  ·  Peltola 84 (47%)

[STAR Voting: Winner — STAR Voting Method (single winner)]
 Begich

And the RCV-IRV rounds show the squeeze and the exhaustion directly:

ROUND 1
Peltola  80  Hopeful
Palin    63  Hopeful
Begich   57  Rejected      # the Condorcet winner, cut for fewest FIRST choices

FINAL RESULT
Peltola      96  Elected
Palin        92  Rejected
Blank Votes  12  Rejected  # ballots that exhausted when Begich was eliminated

Peltola's 96 is only 48% of the 200 ballots cast — a "majority" of the surviving ballots, on a denominator shrunk by exhaustion (the paper's majoritarian failure).

Full engine detail: STAR report · RCV-IRV report · run it: bv2213_k3fmwv_alaska_2022.yaml.

The same profile also carries the other IRV pathologies

The paper documents several failures in this one election; the reduced model reproduces each:

  • Spoiler. Remove the losing candidate Palin and Begich wins — a candidate who couldn't win changed who did.
  • Upward monotonicity. Had ~6,000 Palin-bullet voters (≈6 in the model) instead ranked the winner Peltola first, Peltola would have lost: those votes eliminate Palin first, and Begich then beats Peltola. Ranking the winner higher defeats the winner.
  • No-show / truncation. Had ~5,400 Palin › Begich voters stayed home, Palin is eliminated first, Begich beats Peltola, and those voters get a better result by not voting.
  • Majoritarian failure. The winner holds only 48.4% of all ballots (48% here) once exhausted ballots are set aside.

Methods that read the whole ballot — Ranked Robin, any Condorcet method, and STAR via its runoff — sidestep all of these, because none of them gate survival on first-choice counts. The lesson isn't "ranked ballots are bad"; it's that the tabulation matters, and Hare elimination is the part that failed here.