There's a difference between riding on hard pack with a bike, and driving a car over it. I have no idea why you think an experiment with hard-pack snow on automotive roads is in anyway indicative of what works on cycle lanes.