Literally running into the exact same issues as you did. Had the same results when translating by hand, and same results when using the Kmeans with Naive Sharding Centroid Initialization. I appreciate that this challenge led me to learning new things about clusters, but I came here to improve my coding skills, not my statistics skills, and I am beyond frustrated at the moment.

Update: I was able to get my solution to pass. Thanks for mentioning the initialization, as that is where mine was seemingly off.

I am not sure if there are easier ways to describe an algorithm of a solution of a 2 kyu kata.