Ian,
I'm totally and fully in line with your points. Stability and reliability are the two main points, far before even features. Smoothness in upgrade process is another critical point. So far Zentyal doesn't achieve any of these thus if it matters for you, you have to find one version fitting 100% of your needs and decide to never update. This is what you did with 1.4 and what I do with 2.2
On the other hand, I do understand that there is a need for additional features from time to time and some pressure from potential customers (well, rather potential community users) asking for the very last tool and pushing because another similar product is able to provide something that is missing on Zentyal side
There is a real competition here
Because of these, the idea that is to align Zentyal development with Ubuntu release makes some sense assuming, if looking for stability, you decide no to follow this for your servers in prod.
That, to me, where the drawback is: testing server on testbed, meaning not in prod as very little added value because you may not notice some problems and will not be able to measure if this new release fits your expectations in a day to day activity. This is where I'm in line with Robb but I'm afraid there is no middle way. I would say that current release policy will mainly benefit to Zentyal customers and also to community users who are able to resist to the very last release announcement but will also minimize efficiency of feedbacks from community.