Can you win most of the seats without winning most of the votes?
The same 36 voters. Three district maps. Three different outcomes.
Same voters, three different winners: whoever draws the lines can pick the result. The same voters produced a 2–2 split, a 3–1 Red result, and a 4–0 Blue sweep.
The same is true of school attendance zones, sales territories, and any contest decided district by district: the boundaries can matter as much as the votes.
Who draws the lines in the contests that matter to you?
How can boundaries change the result?
District elections count separate groups rather than one overall electorate. Moving the boundaries changes which voters are counted together, so the same total vote can produce different numbers of seats.
See the same principle somewhere unexpected: The 1960 World Series shows how winning the units that count can matter more than the overall total.
Could a real congressional district look like this?
Yes. These three actual U.S. congressional districts are drawn to their real boundaries.
Illinois 4th — “The Earmuffs”
Two separate arcs of Chicago — one largely Puerto Rican, one heavily Mexican-American — were joined in a single district by a narrow connecting corridor.
Maryland 3rd
One of the least compact congressional districts of its era, its boundary looped through disconnected-looking pieces of the Baltimore–Washington region before being substantially redrawn after 2020.
North Carolina 12th
A long, narrow district following the I-85 corridor. This version became part of the racial-gerrymandering case Cooper v. Harris (2017), in which the Supreme Court invalidated the district.
The strange shapes are real, but shape alone doesn’t explain why they exist. An unusual boundary can reflect partisan strategy, minority representation, voting-rights law, geography, or several forces at once.
Related exhibits
Sources: The 36 voters and three grid maps are invented to show the mechanism. Real districts: 113th Congress congressional district boundaries (2013–2015), U.S. Census data via the unitedstates/districts open-data project. Historical context includes congressional and court records, including Cooper v. Harris (2017).