A colleague at work forwarded this clinical trial to me, and I had to share it with the community here. Not much to add, just read, make sure you read some of the comments too.
https://www.bmj.com/content/341/bmj.c6801

Coincidentally, I just ran across that today, too. I thought the comment on the perils of surrogate outcomes was particularly incisive.

Nice! The results aren't surprising, given the variables encountered in commuting. I suspect that he'd have seen similarly random results if he'd isolated any of a range of similar variables---which songs were repeating in an irritating loop in his head on which days, etc. In fairness, similar trials were conducted on a somewhat broader scale in professional racing in the 1990s, with different results.

He didn't really figure out anything most of us that ride don't already know.

First, the delta between a 20lb and 30lb bike doesn't matter much once you add in the rider weight and other stuff you're carrying.

Commuting time is dominated by traffic control if you ride legally.

If he went for 50 - 60 mile rides with his friends through the countryside on the weekends at 16 - 20 mph, his experience would be greatly different.

At lower average speeds with constant stop and go, it's more important for the bike to be adjusted and well lubricated.