It does not really matter whether the mutations occur on fast or slow markers. The difference is small compared to the uncertainty in mutation rates. Counterintuitively, a mutation on a slow marker indicates a closer relationship than one on a fast marker
Given the otherwise close match, I would count it as 1, since it likely occurred in one (maybe, 2) event. However, it is hard to come up with a mechanism. I am a little skeptical of FTDNA's statement and would recommend their taking the DYS464X test, as well as testing other descendents.
However, they do match closely enough for the DNA to be consistent with paper trail. That is all you can really get from testing.