User blog comment:Googleaarex/Pair sequence number analysis/@comment-27516045-20170725134920/@comment-30754445-20170727100934

I don't think it can.

But more interesting is the fact that this very same sequence provides a counter-example to my claim on the other thread that "BM1 and BM2 always result in the same 'bad parts' and the only difference is the C-matrix".

for (0,0)(1,1)(2,0)(1,1)(2,1)(3,1) they don't.

I guess this means that I've interperted the source code wrong. I'll try to look at it again later.