User blog comment:Koteitan/Categorizing of the rule sets for all sub versions of bashicu matrix/@comment-11227630-20191122094441/@comment-35470197-20191124090602

> Exactly. That's strange point of BM2.

But according to Bashicu's slide, Bashicu determined the bad root by referring to all rows, as I wrote above. Is Bashicu's slide inconsistent with the orginal definition?