At the moment it is very easy to over- or under- plan high level capacity by having releases overlap, or by having big gaps amongst releases. This is very dangerous, and either results in unrealistic expectations being set to stakeholders and/or capacity planning at this level being ignored. Aha! doesn't give any feedback if you're wildly over or under committed; it relies upon you having calculated that the capacity and dates of a release are consistent with each other, and that you have checked you don't have big gaps amongst releases.
By pulling the concept of capacity out of releases and putting it into (a new concept called) "planning increments", you would have this feedback. You can continue to plan out a release and define the features that make it up, without the danger of being wildly off the mark in terms of capacity. The Feature board would have the releases running horizontally as rows, with the planning increments as the columns. You can then drag features across planning increments and see the usage of capacity updated. Releases could then have their dates set automatically according to the date of its first and last features' planning increment, with the potential for Releases of different initiatives be run in parallel while being confident that the plan is still realistic. This would bring a release back to its raison d'etre: MVP; through the features that make it up, which can then be planned across the roadmap.