Speaking of two-set blocks, I've actually been a fan of the current structure. The old model of a core set, followed by a three-set block, has its pros: if you land on a great block idea, you get to really develop its themes over multiple sets. But it is very prone to boredom upon the last set.
On the other hand, one set blocks would barely touch upon the themes and mechanics the world has to offer. Imagine if Ravnica was only given one set, or Time Spiral. What a waste!
The current two-set blocks seem like a nice middle ground. Plus, with two blocks per year, there is no room for core sets (yay!). But the two-set structure can still suffer from blocks that wanted a third set, or blocks that only needed one.
Perhaps these issues could be solved by frequent returns to recent planes. For example, you could have a set based on a new plane, then the next set be a smaller standalone, and following that a return to the first plane. For example, instead of:
AKH (large set) -> HOD (small set) -> XLN (large set)
You could break up the Amonkhet sets with a smaller Ixalan set, like so:
AKH (large set) -> XLN (small set) -> HOD (large set)
This would be a neat way to give deeper blocks like Amonkhet the cards/sets they deserve, without boring players with multiple sets of a block that has less to offer (such as Ixalan, IMO).
Hopefully Wizards will eventually land on a similar model after experimenting with single set blocks. Overall, I feel that block design needs to be a lot more flexible and dynamic.
