[LexxM3]

I am idly wondering how such an unquestionably idiotic design managed to make it into a production car. It not like this was an oversight or an accident, something as critical as a mechanical transmission unlock is not going to just slip under the radar of the many DFx (x=Serviceability here) reviews in a robust engineering process. So I don't think it was an accident or a result of an oversight, I think it was a "classic" misguided engineering trade off. Here is my best imagined guess how it went ...

The starting points are that a mechanical unlock in a powerful high torque AWD car is already a difficult issue and that all previous designs somehow routed a mechanical override inside the cabin (slot for a key). The inside-cabin design may come with significant mechanical complexity and increase in complexity is always a reduction in net reliability, perhaps to the point that, PROBABILISTICALLY, adding that complexity can obliterate the serviceability benefit. These decisions are very complicated and never perfect and so BMW (and many other serious engineering teams) use reliability simulation software to help with tradeoff decisions. I strongly suspect that's how that happened — the simulation said this approach was by far more net reliable in the overall design structure of the platform (that needed to support pure ICE AND hybrid PHEV-with-batteries drive trains).

But here is the Achilles heel and why rational thinking is still necessary — the simulation is based on thousands to tens-of-thousands of assumptions and can be effectively "unstable" (random?) due to sensitivity to those assumptions. It's a quantitative and theoretically "objective" tool, but engineers still need to apply careful judgement to the tool's input assumptions and output results. So I am not letting the BMW engineering team off the hook, but it is interesting to analyze how design (and actually all) failure occurs.