Sandia also did some work on improving the savonius. They came up with a revised blade shape that gets just a squidge under 2/3 of Betz.
(So make it 23% bigger in each dimension and get more power than a darrieus could without the materials and metal fatigue issues. Savonius rotors are strong and can be made out of sheet material - flat and bended - which is inexpensive and easy to form and join.)
Yes they did a bunch of work with the eggbeater-style darrieus, and didn't get to a really practical design. But the latter is not too surprising: Darrieus style mills have forces that change direction radically every rotation. You also have turbulence when the blades are downwind of other blades or the hub. And NASA did a bunch of work on the blade profile so it would stall-furl and produce a flat power output (at the design power for the genny/transmission) for a range of wind speeds, which throws a bunch MORE turbulence at the downwind blades - a price worth paying to keep the high TSR blades from trying to go supersonic in a storm.
With thin blades and strong oscillating forces you flex the metal. NASA decided to build it out of aluminum to keep the weight down. But aluminum has terrible problems with metal fatigue when it flexes, stress-hardening at a small fraction of its yield strength. So a design like theirs, as with an aluminum-frame aircraft, has a strictly limited life - and if you don't disassemble it once it gets elderly it will do it for you.
A savonius-style rotor can easily be made out of steel (which doesn't start stress-hardening until it's flexing up to a goodly fraction of its yield strength) and the only major penalty from the higher weight is that you need bigger thrust bearings. With a TSR a bit less than 1 you won't have any parts whipping around at 600 MPH in a storm and you can build it strong enough that the rotor will survive without furling. AND it's self-starting. B-)