Alaska 2022 — a real RCV-IRV center squeeze, spoiler, and monotonicity failure¶
Alaska's August 2022 U.S. House special election is the most-cited real-world example of RCV-IRV misfiring — because it hit three of IRV's pathologies at once, in a high-profile federal race. It's the case advocates on both sides keep pointing to.
Which Alaska election is this? The August 2022 US House special — the one that misfired. Alaska has run many other RCV races (the Nov 2022 general, the 2022 Senate, 2024) that worked fine. For the full map and honest scorecard, see Alaska's RCV elections — the whole picture.
→ Run it yourself: Alaska 2022 — reduced 200-voter model, four counts (BV-backed; STAR & Ranked Robin elect the Condorcet winner IRV cut) · Interactive breakdown (Equal Vote): realrcv.equal.vote/alaska22 · concepts: center squeeze · non-monotonicity · the spoiler effect
What happened¶
Three candidates, one seat, counted by RCV-IRV (Hare):
- Mary Peltola (Democrat)
- Sarah Palin (Republican)
- Nick Begich (Republican)
The count, round by round (certified results):
| Round | Peltola | Palin | Begich |
|---|---|---|---|
| First choices | 40.2% | 31.3% | 28.5% ← fewest, eliminated |
| After Begich's ballots transfer | 51.5% ✅ | 48.5% | — |
Begich had the fewest first choices, so RCV-IRV eliminated him. His ballots split roughly half to Palin, under a third to Peltola, and about a fifth exhausted (those voters ranked only Begich). Peltola won the runoff over Palin.
Why it's a teaching case: three failures in one election¶
1. Center squeeze — the Condorcet winner was eliminated. Head-to-head, Begich beat both Peltola and Palin — he was the Condorcet winner, the candidate a majority preferred over each rival one-on-one. But because he was too few voters' first choice, IRV cut him in the first round. That's the textbook center squeeze: the broadly-acceptable middle candidate is knocked out for lacking first-place "core support," even though he'd have won every direct matchup.
2. Spoiler — Palin spoiled the race for Begich. Had Palin not run, Begich would have won (his support plus Palin's beats Peltola). Palin couldn't win, but her presence changed who did — the definition of a spoiler. So RCV-IRV did not deliver the spoiler-free election it's marketed as.
3. Non-monotonicity / favorite betrayal. The result was also a documented monotonicity failure: a bloc of Palin-first voters would have gotten a preferred outcome (Begich instead of Peltola) by ranking Peltola higher — i.e. by not putting their favorite first. IRV punished sincere voting, exactly the favorite-betrayal incentive reform is supposed to remove.
Why it matters¶
Alaska 2022 isn't a contrived textbook ballot — it's a real federal election where the winner was not the candidate a majority preferred head-to-head, and where a losing candidate flipped the result. Methods that read the whole ballot handle this cleanly: Ranked Robin (and any Condorcet method) would have elected Begich, and STAR resolves the squeeze through its automatic runoff rather than by first-choice elimination. The lesson isn't "ranked ballots are bad" — it's that the tabulation matters, and Hare elimination is the part that failed here.
For a follow-on: after 2022, party elites pressured a Republican to drop out of the 2024 race to avoid a repeat — the real-world "exit incentive" a spoiler-prone method creates (see origins & spread).
Sources: Alaska 2022 — Equal Vote realrcv, 2022 Alaska's at-large special election — Wikipedia, Graham-Squire & McCune, "An Examination of the 2022 Alaska Special Election"