Am I right in understanding that every time a rider has a bike problem, you're blaming it on Di2, whether or not it was a Di2 problem, and whether or not the rider was even using Di2, and then using this as evidence against using Di2?
that would be too easy. Little clues like the rider being stuck in the wrong gear entirely, on a bike with no gear cables tend to give the game away....
one of the many incidents I recall was a few years ago when the ToB went through Keswick. There were two riders in a chasing group and one of them went for another gear at the back. P-twang went his electric RD, right into the spokes. This was a race-altering event and I saw it on live TV, but it didn't make the highlights show despite being pivotal to the events of the day's racing.
cheers
My personal opinion is that the idea of relying on a spring to pull a cable back through its housing is fundamentally flawed for any sort of bike that's used by a non-pro - i.e. used in all weathers and not necessarily serviced literally after every ride. It may work well for lots of people but I don't consider it reliable enough.
I've personally written off the idea of a cable actuated derailleur for any bike on which I do any sort of serious event due to empirical observation, from two incidents.
One being PBP in which the shifter cable became frayed. Luckily the mechanic at Brest was able to change it but if I hadn't had that luck it could have been ride ending - he said it probably had another 100km until the few remaining strands snapped through.
Second was on a commute when a gunged up cable was too much friction for the shifter, and the shifter itself (SRAM) snapped.
I am happy to trust the rohloff shifter which is a sealed system, both cables are hand actuated and the same thickness as a normal bike's brake cable (in fact the inner
is brake cable). I'd also be happy to trust Di2. But I personally think the idea of relying on a spring to pull the cable back through is fundamentally flawed. The knock on effects of the spring not being strong enough are that you have to make the cable thin and reduce the amount of outer to reduce friction as much as you can, which leads to problems with ingress.
I only use a cable derailleur on the utility bike but it seems to work well on that I think largely because it has only got 6 sprockets, I rarely ride it in the wet, and failure isn't as much of an annoyance as it would be on an audax.