Actually they're 1.17x. The problem is that the Rinard method can only be used on ONE flange. The other flange requires a different pattern due to the lack of proper flange symmetry.

Unless someone else can think of something, as far as I can tell, there's nothing special about 2.17x that compensates for this.