Because it has less to do with the rider, and more to do with the cycle itself, than previously thought.
This article from Ars Technica details a study where a bunch of scientists, simply put built an ‘unstable’ cycle. A cycle that defied all the hitherto-postulated theories for the tremendous stability of a moving cycle. (Ever tried to cycle without hands? It’s easy).
A cycle that still refused to fall over.
To test the relative contributions of these factors, the authors eventually built their own computer model of a bicycle and started playing around with various features. It turned out that they could eliminate both the gyroscopic and the negative trail factors, and the bike would still be stable as long as it was moving faster than 2.3 meters (7.5 feet) per second. They could even move steering to the rear wheel and produce a stable design.
The apparently unreasonable stability of different bicycle designs must have suggested that their model had probably lost touch with reality, so the authors went out and built a bike with a counter-rotating wheel to get rid of gyroscopic effects, as well as a negligible (4mm) trailing between the front wheel and the steering. As their model predicted, it tended to stay upright, and would steer into any falls that their grad students tried to induce.
As someone who’s spent a LOT of time on a cycle and in the pool, here’s my armchair-physics theory:
A bicycle’s stability is based on its center of gravity being under the bike. Move the CoG out from underneath it, i.e. – tilt it, and it topples.
A cycle is stable a lot in the same fashion a pro swimmer does not wobble and twist when swimming. While the latter is a result of training, the former (imho) is a result of no more than a simple and rigid design. Take away the rider, and there’s almost nothing left to influence the the CoG of the bike. The only ‘unstable’ part of the bike is the front wheel, which again is akin to the head of a swimmer, and ‘leads’ the entire body in a direction rather than destabilizing the system.
I’m sure real physicists could do math and prove (or more likely, debunk) my thoughts, but such is as they are.