Wheels on racing vehicles are as small as possible because rotating mass (and unsprung weight) is a huge issue. Racing vehicles will nonetheless run very big wheels at times, because they need to fit very big brakes-- not an issue with bicycles. Bikes are fairly limited as to what their tire height can be relative to width-- cars do not have this limitation, because they can run a tire with 5 plies in the 6" high sidewall.
I agree with everything stated about hysteresis, but there is obviously a practical limit-- if there wasn't, we'd all run motorcycle-sized tires... and there's Fatbikes, which do pretty much that, but are inarguably niche bikes with a narrow range of applications. This is (IMO) why their test seems to completely ignore the flywheel effect-- if they had tested a solid aluminum 20" wheel with any BMX tire on it, it would roll farther than all of them. End of the day, once fitted with an appropriate tire, 700c, 650b, and 26" all stand roughly the same height-- so it's not surprising that they would all roll similar distances. Somewhere in there is a point of diminishing returns.
__________________
