C4D has advantages of stability, intuitiveness, ease of creating models, simplicity (of things like mograph), NPR rendering, fur/hair, etc. For things like modeling a human, I would use this (or Modo or Maya or 3Ds Max) any day over Houdini, as Houdini is going to save every edit as a node. Where Houdini excels are things like particles, fire, fluids, etc. Its parametric paradigm means that one can substitute complex models for simple objects after the fact. One can do a bit of this with cloners in C4D, changing spheres to teacups etc., but unlike C4D the substitution is much more flexible and not limited to some subset of the program. For example, one can create tentacles of an octopus with something like a sweep nurb, but in Houdini, one can then select certain polygons and create bevels like the suction cup parts. The original curves are still there, so they can be animated, yet the sweep is much more complicated than C4D allows without converting to a polygon object.
In addition, points can have an infinite number of attributes, not just position in space over time. One can create an attribute that is based on some function over time (like sin(time)) or add an attribute of some randomness between 0 and 1. These additional attributes can be used to drive models, particles, etc. Unlike C4D, you seldom paint yourself into a box. For instance, (excluding the bevel effector in C4D), if one creates a complicated bevel and wants to reduce the number of polygons, it just requires a click in Houdini. One has to do a lot of remodeling in C4D. The rendering system (Mantra) in Houdini also allows much more complicated shaders. However, all this comes at an expense of complexity. For instance, creating a spring (sweep nurb of circle along a helix) is a couple of clicks in C4D but requires several not completely intuitive nodes in Houdini.