News for hackers. Stuff that matters.
Re:How to get rid of gerrymandering: the math is surprising by JJMcJ on Thursday May 12, 2022 @02:52PM (#31358581)
Square state, two Congressional districts, relatively uniform population distribution but one party clustered in one section of state.

Sometimes you get different outcomes if you divide state in half north-south vs. east-west, depending on how the cluster is divided between districts.

Re:How to get rid of gerrymandering: the math is surprising by lkrubner on Thursday May 12, 2022 @04:55PM (#31360027)
Are you describing the state you live in?
Re:How to get rid of gerrymandering: the math is surprising by JJMcJ on Thursday May 12, 2022 @08:03PM (#31361579)
No, I once saw this described for Nebraska. I don't remember the details or which party benefited from each possibility.

Here's an idealized example. Imagine a state with 1,000,000 people. 400,000 Democrats live in a narrow strip on the Eastern border, uniformly distributed in that strip. The 600,000 Republicans are uniformly distributed in the rest of the state.

Split the population in half by a north-south line. The Eastern district has 400,000 Democrats and 100,000 Republicans, and so has a Democratic representative. The Western district has 500,000 Republicans and so had a Republican representative.

Now try an east-west line. Each district has 200,000 Democrats and 300,000 Republicans. There are two Republican representatives.

Yet on a map, both divisions look fine.

That's one reason the anti-gerrymandering rules specify looking deeper than just population and compactness.

Re:How to get rid of gerrymandering: the math is surprising by lkrubner on Friday May 13, 2022 @11:13AM (#31368703)
Yes, good example.