This article makes the fair counterpoint that just taking a thing and throwing it into a random test doesn't make for very informative data.
theconversation.com/what-is-a-…
A propeller or turbine will be optimized for specific circumstances and to say if toroidal design is better in some way to classic design, you'd have to build an optimized model of each for some specific scenario of whether it's pushing or being pushed, what the medium is, what the load is and what the speed is.
So maybe I'll have to read those papers after all. 😊