Each state’s Electoral College votes equal its House seats plus its two senators. The House grew after almost every census until the Apportionment Act of 1911 set it at 435 seats, and the Reapportionment Act of 1929 made that permanent. Every presidential election since 1912 has been decided with a 435-seat House.
I wanted to know whether that cap ever changed who won, so I built an Electoral College simulator that replays every election from 1912 to 2024 with a different House size.
What it does
- Compares House size rules. Every election under the real House, the cube root rule, and the Wyoming rule, with changed outcomes highlighted.
- Lets you pick the House size. Choose a fixed size, a number of people per representative, or a custom size after each census, along with the apportionment method (Huntington–Hill, Webster, Jefferson, or Hamilton).
- Finds the flips. For any election, it tries every House size from 100 to 2,000 seats and charts where the winner changes.
- Handles split states. Maine’s and Nebraska’s district electors and faithless electors are modeled as closely as the data allows.
What I found
Only one election changes under the cube root rule: 2000 flips to Gore. With a 629-seat House after the 1990 census, Gore would have won 368 electoral votes to Bush’s 364. Under the Wyoming rule, no election changes.
2000 is also the election most sensitive to House size. Bush wins with any House of up to 490 seats and Gore wins with any House of 656 seats or more. In between, the winner changes back and forth more than 50 times.
Data and source code
The election and apportionment data come from the U.S. Census Bureau, the National Archives, the MIT Election Data and Science Lab, and Wikipedia. The simulator reproduces every state’s real electoral votes in all 29 elections and every real apportionment since 1940.
The code and data are open source on GitHub.