That'd give an upper bound of 243.8, which is far too close to 240 to be practical. As I tried to add, there is some necessary and natural deviation from nominal to account for path loss etc. 230±6% only allows 240 to deviate by 1.5%, which isn't much scope at all.

I believe ±10% was the traditional definition, so they just trimmed a little off the top end so it strayed less from Europe's definition - but not enough to actually affect nominal distribution.