Forced Exhaustion Ceiling — when a ranking cap discards more ballots than the winner gets¶
Method: RCV-IRV (Instant Runoff) · 1 seat · Expected winner: Ada · full count →
▶ Live on BetterVoting: vote · results ↗ (BV2183, election dfw8rj).
Read this fairly. This is a deliberately constructed worst case, built to show the ceiling of RCV-IRV ballot exhaustion under a ranking cap — not a claim that real elections look like this. Real exhaustion runs milder (roughly 10–27%; see the real rates). The point is what the mechanism permits, and that the cause is the ballot cap, not the voters. It sits in
paradoxes_and_whoops/under the same reading-these-fairly discipline as the Felsenthal/Nurmi cases.
The setup¶
50 voters, five candidates — but the ballot caps every voter at 2 rankings (like a tight ranking limit; NYC caps at 5, Minneapolis at 3). Three minor candidates (Cleo, Dev, Eli) form a rotating bloc; their voters spend both of their allowed ranks on minor candidates.
15:Ada>Cleo ← two real contenders, Ada & Ben, lead on first choices
14:Ben>Dev
8:Cleo>Eli ← 21 voters rank ONLY minor candidates —
7:Dev>Cleo all their ballots will exhaust
6:Eli>Dev
What happens¶
ROUND 1 Ada 15 | Ben 14 | Cleo 8 | Dev 7 | Eli 6 → Eli eliminated (Eli>Dev transfers)
ROUND 2 Ada 15 | Ben 14 | Dev 13 | Cleo 8 → Cleo eliminated; its 8 ballots EXHAUST
ROUND 3 Ada 15 | Ben 14 | Dev 13 → Dev eliminated; its 13 ballots EXHAUST
FINAL Ada 15 (Elected) | Ben 14 | Blank Votes 21
Winner: Ada, 15 to 14 — a margin of one. And 21 ballots (42%) were exhausted — more than the winner's own 15 votes. Ada's "majority" (15 of the 29 still active) is really 30% of the 50 voters. The margin is 1; the exhaustion is 21.
The kicker: lift the 2-rank cap — let those 21 voters rank all five candidates — and in single-winner IRV no ballot can exhaust at all, because there is always a continuing candidate. The exhaustion here is manufactured entirely by the ballot design, not by anything the voters did. That's forced exhaustion, the involuntary kind — the voters were as expressive as the rules allowed.
BetterVoting confirms it¶
BV's own IRV tabulation agrees with the LH engine to the number: winner Ada, and exhaustedVoteCount = [0, 0, 8, 21] — 8 exhausted after round 2, 21 by the final round. Frozen export: …_bv_export.json.
The LH tabulation¶
--- RCV / Instant-Runoff Voting (single winner) ---
Tabulating 50 ballots (ranked ballots).
ROUND 1 Ada 15 | Ben 14 | Cleo 8 | Dev 7 | Eli 6(Rejected)
ROUND 2 Ada 15 | Ben 14 | Dev 13 | Cleo 8(Rejected)
ROUND 3 Ada 15 | Ben 14 | Dev 13(Rejected) | Blank Votes 8
FINAL Ada 15 (Elected) | Ben 14 | Blank Votes 21
Winner — RCV / Instant-Runoff Voting: Ada
Full report: …_tabulated.txt.
Related¶
- Forced vs. Voluntary Exhaustion — the concept this case makes concrete
- Exhausted (Inactive) Ballots — the full picture, including the real election rates
- Ranked Robin — reads every ballot in every pair, so it never exhausts
- Reading these paradoxes fairly