The last 30 years has seen several paradigm shifts in planet formation. Two important concepts deserve attention:

(1) A planet need not have formed at its current location. There is a growing consensus that both inner and outer migration occurs during accretion.

(2) The reservoir of potential building blocks need not be local. A strong case can be made that the Earth formed from planetesimals originating in at least five distinct regions of the solar system: The inner system (out to Mars), the inner belt, the outer belt, the Jupiter-Saturn region and the Uranus-Neptune region.

Just 30 years ago we would have described the outer planets as four ‘Gas Giants’. Today we describe the outer system as two gas giants (Jupiter-Saturn) and two ice giants (Uranus-Neptune).

http://blueox.uoregon.edu/~courses/B...3/FG13_010.jpg

The superficial appearance can be deceptive. For example, compare Jupiter to Uranus in the following pic.

http://zebu.uoregon.edu/~js/ast121/i...ian_worlds.gif

Although it is obvious that there is a difference between their masses, how great is that difference? It’s actually tremendous. Uranus is about 14.6 Earth masses whereas Jupiter is 318 Earth masses.

http://csep10.phys.utk.edu/astr161/l...lanet_mass.gif

IMO, we’ve come a long way. The current models could probably use a little tweaking and refining, but we’re getting there. (Death, taxes and error bars)

This is what I think took place. The cores of Uranus-Neptune formed in the Jupiter-Saturn region, then migrated outward, acquiring mass along the way while simultaneously scattering the objects which they did not absorb.

Note; the following link is 33 pages. Allow a minute or so to load:

http://arxiv.org/PS_cache/astro-ph/pdf/9902/9902370.pdf