BTW it was asked upthread why a Di2 system might be more likely to go into the wheel than a mechanical one, especially given that both have mechanical stop screws.
Well in a mechanical system when you go into bottom gear it isn't unusual for the mech position to be mainly determined by the stop screws rather than the indexing; the elasticity in the cable pulls the mech against the stop screw. Not sure if the same protocol applies to Di2 or not.
What anyone who has worked on many bikes may have noticed is that (as well as bent/badly adjusted mechs going into the wheel instantly) in some cases the rider habitually really hammers the mech into the stop screws and this causes premature wear; the stop screw just digs itself a hole in the part it bears against.
I'd suppose that the same thing might happen with a Di2 system. There is a current sensor that backs the power off if the stall current is reached but until that kicks in the load in the mech is at least as much (actually a lot more) than even the most ham-fisted rider might generate in a mechanical system.
In most of the gears the Di2 system works like any other servo; the system powers until a resolving element of some kind tells the system it is in the right place and the motor quits. Thus a failure of the resolver (or the programming or the adjustment) might allow the system to power into the end stop for bottom gear and appear to work 'normally' but is actually wearing the stop screw at high speed.
Most systems of this type have a feedback arrangement, i.e. they provide a bigger signal to the motor when there is a greater error in position than is normal; thus if the shift is delayed for whatever reason, the motor will exert more torque than normal and may even overshoot the intended position before settling in on it. This could easily damage the stop screw.
With a mechanical system a rider might notice a noise from the mech tickling the wheel during the overshift (i.e. while they are still pushing on the lever) e.g. because the stop screw is wearing, and realise that the system is getting out of adjustment. They certainly won't be likely to keep hammering on the shift lever after the shift has gone in. I'm not sure that the same thing will happen with a Di2 system; if the stop screw limits mech position then it will just wind into the wheel ever more with every shift. If the stop screw is worn or out of adjustment the resolver may position the mech correctly most of the time but if a tiny fault develops in the resolver (or the system doesn't shift soon enough, under load, and then overshoots when the shift happens) then the mech might easily end up in the wheel even if the system worked OK on the workstand.
Remember I said that complex systems have multiple failure modes? Well you might suffer the same kind of fault many times over before being sure what actually caused it; investigating the mangled remains of the mech might not tell you that much....
I guess there are two main things so far as users are concerned;
1) you either do (K) or do not (I) understand how the technology works and
2) you either do (B) or do not (U) believe that the technology is implemented in such a way as it is less likely to go wrong/be unfixable vs alternative systems.
I guess if anything I'm a 'KU'; there are a lot of folk who are IB and some who are KB.
cheers