The Mudroom 🥾 — a criterion-failure gallery (deliberately one-sided, openly so)¶
What this is, honestly. Arrow and Gibbard–Satterthwaite prove that every voting method must fail some reasonable criterion — there is no perfect method, and there never will be. The Mudroom is the empirical companion to those impossibility theorems: a per-method collection of concrete, runnable examples of each method on its worst day. The twist is that each reel is written unfairly on purpose — curated with zero charity, the way an opponent would — because seeing the mud slung at every method (this repo's favorites included) is what makes the fair pages trustworthy. The educational core is real; the gleeful tone is the indulgence.
Ground rules for the whole room: 1. Every example is real and documented straight (fair, hedged, rarity-stated) elsewhere in the repo. The curation is the only unfair thing. 2. Every reel links its own conscience clause and the grown-up version you'd actually cite in a debate. 3. No method is spared — that's the point. If a reel for a method we like looks thin, that's a to-do, not a verdict.
Plain warning, up front: these pages are deliberately mean, unfair, and one-sided. They are not balanced and are not for citing. They're kept low-profile and clearly labeled for exactly that reason — not hidden, just not featured.
The reels¶
| Method | Reel | The headline whoops |
|---|---|---|
| RCV-IRV | The Non-Monotonicity Files | raise the winner → they lose; real cases (Burlington, Alaska, San Francisco) |
| STAR | The STAR Whoops Files | not Condorcet-compliant; fails LNH / majority / mono-raise-delete; burial squeezes the center |
| Ranked Robin | The Ranked Robin Whoops Files | cycles (no winner to find); intensity-blind sincere dark horse; tie-by-lot |
| Approval | to fill as we find stuff | the chicken/Burr dilemma; the "where do you draw the line?" threshold trap |
| Plurality / Choose-One | to fill | the spoiler effect; wasted votes; can elect the Condorcet loser |
| Score / Range | to fill | reduces toward Approval under strategy; needs a fixed common scale |
| Borda | to fill | the Dark Horse — a nobody wins with zero honest support |
(The empty rows are honest to-dos — a method missing from the room isn't innocent, just un-curated yet.)
Why it's low-profile (not hidden)¶
These pages are mean, unfair, and one-sided by design — they'd misrepresent the state of the art if anyone read one thinking it was balanced. So we don't hide them (nothing here is secret, and it stays in the published site), we just keep them low-profile and unmistakably labeled: off the main navigation, named with a leading underscore, and every one opens with a plain "this is deliberately mean and unfair" warning. Findable if you go looking or follow the one tucked-away link from paradoxes & whoopses; never presented as the honest picture. The balanced, cite-in-a-debate versions live where they should — criteria at a glance, each method's honest-limits page, and how to read these fairly.
Where this sits — the "Level 401" safety check¶
Think of the library as a course. 101 teaches how a method works (fill out the ballot, run the count). 201 / 301 teach its properties — pros, cons, the criteria it meets. The Mudroom is the unofficial "401": the safety check. Once you understand a method and like it, the last responsible step before you'd adopt it is to ask "how does this go bad?" — to walk its failure modes with clear eyes. Every method has whoops; a grown-up chooses one knowing exactly which.
The fair version of that 401 already runs through the repo — each method's honest-limits page, criteria at a glance, the four-part test. The Mudroom is the same safety-check material with the bedside manner removed: the failure modes curated one-sidedly, for sport. Read it for the reels; do the actual 401 on the fair pages.
The actual lesson (if you want one)¶
Strip the mud off and the Mudroom teaches the single most important fact in voting theory: there is no perfect method. Every ballot design and every count trades something away — the honest question isn't "which method is flawless?" (none is) but "which failures can you live with, and how often do they actually bite?" That's the four-part test and the whole reason this library exists. The reels just make the impossibility concrete — and, by refusing to spare anyone, keep the rest of the repo honest.