User blog comment:Googleaarex/SGH Functions/@comment-25418284-20130404163840/@comment-5150073-20130406230546

It turns out that just replacing all w's to n's is not always correct in SGH. If it is true, then SGH should catch FGH at large Veblen ordinal. However, it has been proven that g_BHO(n) ~ f_e_0(n).