I have never not seen my dancing blade register as a true combo in training mode personally, and I feel absolutely no change in my dancing blades. I have been killed plenty of times by marths up combo of his dancing blade if a tipper lands at the percents you mentioned, so I am still a bit skeptical. I am currently using my dancing blade on greninja as lucina and the combos you mentioned seem exactly the same to me. Without rage he isn't dying at 110% to the forward variation, but I know with a little bit of rage he will, just its always been. On Marth I just used the up variation on Greninja on FD and he didnt die off the top at 145% without a tipper, but when I sweet spotted it he died at 110%, which seems about the same as it has always been for me.
This is always the kind of stuff I am skeptical about though, there are a million different reports coming out of things "feeling different" just because someone doesn't have complete knowledge of a character and then that mentality spreads to others thinking that they feel it too. There are so many factors that determine when a move can kill but people will come out and be like "oh, this kills 10% earlier than it used to. Without hard evidence or the move is drastically changed, it is damn near impossible to tell in a situation like this. Dancing blade is basically the bread and butter of my Lucina and Marth game play, and I am still leaning towards it being unchanged personally until the data shows otherwise.
Lucina