Something folks need to remember is that designing features for turbulent flow is a science that's still in the developmental stages. They've only been able to do computer simulations of turbulent flow, vortices, eddies, etc. for about a decade, and I'd expect it's pretty expensive to develop a computer model of a whitewater park so they can simulate features at all flows. It's a lot easier to say, after the fact, that someone did a crappy job because the feature didn't perform well during certain flow conditions, than it is to design a feature that'll work well in all flow conditions.
Does the feature work as needed 95% of the time?
Nothing in the world is more yielding and gentle than water. Yet it has no equal for conquering the resistant and tough. The flexible can overcome the unbending; the soft can overcome the hard. - Lao Tse