Interesting. Now consider the following scenario, which perhaps led me astray.
Case A) Assume 4 riders all with the same power output and aero losses when riding by themselves. Have them ride 2 across and 2 deep, so the same scenario as tested by the magazine, but with each lead rider having someone behind.
Case B) Same as case A, except there are only 2 riders, with one in front of the other.
I thought that case A should be faster than case B. And a case with 3 across, each with a lead and follower should be even faster. Is that correct?
No, I don't believe so. The net drag of 2 riders side by side is slightly more than double the net drag of 1 rider, and I don't see why it would be different in a double row.
And the side by side thing is not just a theoretical effect, or one that only shows up under controlled circumstances in a wind tunnel, it is absolutely a verifiable real world effect. It's not news to me because it's something I've known about from a skydiving background. When planning large formation skydives it has to be taken into account that the more people join a set style of formation, the slower it's going to fall. For real.