Scores average every run and implementer. Hover a dot to see whose library it is; click a problem to see each implementer. ± is std across runs.