Wel given we can reverse-engineer all the numbers (as I said before, the three groups all must have had precisely 2 failures each out of one of 19,20 or 21 in each group) I hardly think it matters.
My best guess is they put out the PR, and then weren't comfortable making a decision "on the fly" during the call to release more data than they originally put out. So I'd just put it down to a bit of inexperience PR-wise here.
News 
Market Data 
Discover 
