User blog comment:Googleaarex/BMS encoded to 1 symbol system/@comment-2001:56A:71A8:5F00:65F5:A83D:365D:2232-20180226031537/@comment-30754445-20180226062646

It's a questionable claim, given that:

(1) We aren't even sure that BMS is well-defined much further than the pair-sequence system (that is: that there aren't any infinite loops).

(2) Even if we assume that it is well-defined, none of the analyses of the system I've seen so far bother to give any justification what-so-ever for the claimed results.

So for all we know BMS might fail shortly after the TFS, making BAN a stronger notation.

At the other extreme, if we accept some of the analyses of BMS we've seen here at face value, it might actually be stronger than SAN. The common belief here is that (0,0,0)(1,1,1)(2,2,2)(3,3,3) is way stronger than anything that can be practically analyzed.