Some points to consider, while using only a Falloff based weighting.
Are they child objects of joints or stay stationary? Each situation creates a couple of problems.
Stationary, if the axis of the “character” receives any change, all weightings might be off. If the character is not moving, things stay stable.
If the falloffs are moved with the joints, the total value of each point must be 100%, which is hard to keep, given the rough details that falloffs have compared to the point-based mesh. overlapping values will be handled with an evaluation. This means the mix for each point might undoubtedly change. If the total of all values is below 100%, then points are moving not all the way, which is only in small distances, if at all, tolerable.
Falloffs centered on the joint, will affect the parent joint influence of the points. This will result in a skin that seems to be moved by a broken bone. Which is highly unlikely in regular use.
The information on single-point values is not given. This means no information about how much each falloff/joint contributes is not available.
Any IK set up with Stretch, and Squash options will have no option to change the falloff accordingly. This will lead to a highly unpredictable deformations. Of course, an Xpresso set up might help here, but is the complexity worth it?
The idea of using the Freeze option in the Falloff might come to the rescue of some problems mentioned so far. The question of why not bake the values directly, instead of such a huge management requirement needs to be addressed.
A falloff might touch other areas of the surface that will be affected unintentionally.
The values produced, can’t be stored in a file (Weight manager> File: Save/Load.) This creates a limiting situation, as the set up must stay intact. If that is needed, it might be of minor importance.
There are, without a doubt, even considering the problems, perfect uses for such weighting, that it should not be avoided at all costs. The option to dynamically change the weighting, or have interactively auto-weight overlapping areas, can lead to new and otherwise impossible solutions.
I do not assume that this list is complete, nor a fit to any possible set up, but the use of falloffs should be done with good knowledge about its pros and cons.