How real do you believe some of these CDAs are ?

My hunch is that people are getting lower CDAs but mostly through more availability to testing and they are not going from 0.25 to 0.18.

I was part of the tests they did with Lionel when filming chasing the Lion. The first run was with his 2015 Kona position including the Camelback. He was a .260. Very soon he was 0.24. By the end of the session he was a 0.23. He since moved on to a new bike, new fitter, lots of testing and in his hour record he achieved what ? 0.22 ? (which includes much better tires)

I don't think frames, helmets or wheels have changed much in the last 5 years. Lots of data to show that. Put a 2016P5 with a LGP09, HED disc/90 against any combo today. A watt ? Two ?

I do think fabrics and assembly of suits is improving slightly.

I think rubber has improved quite a bit. I don't think there was much under .0035 (that could last more than 4km) a few years ago.

But I think the biggest gains come from the availability of testing from multiple resources. There are more tools and more knowledgeable people. In 2016 very few riders had been tunnel tested, which was the only way to test. Today, every rider has been tunnel or velodrome tested with some Chunging and creative use of sensors. I'd love to hear a presentation from Dan Bingham on the gymnastics of data/sensors he uses to do his magic.

There are riders (like Sebi) that optimized early and really profited from it. But it has now become the norm. And I think this has driven a lot of the improvements. But soon the "elites" will hit the peak of aero.But the masses have a lot of work to do.

There is still a lot of room for improvement on the test tool front. We are seeing the tip of the iceberg but I think steam is running out for a few of the vendors and I think there is potential for a large paradigm shift in the next 18 months.