It certainly looks like overly complex. The simplest system, that is, airbags, are unusable because of the size (it's almost like the difference between a big toy car and a regular car)
You probably can't brake it only with retrorockets, and as this mission is aiming to be the most precise landing on Mars ever (and I'm not even sorry for the pun) they need a way to get the lander to an exact location (by deploying it from the crane)
People seem to forget/not realize that the EDL (entry, descent and landing) of Spirit and Opportunity had pretty much all the elements that the Curiosity EDL does PLUS the airbags and the airbag cocoon: http://www.youtube.com/watch?v=Ij33yhdGn_g
Leaving out the sky crane means that the comparison doesn't involve "pretty much all the same elements." The MSL sky crane is insanely complicated. Not much can go wrong with an airbag, relatively speaking.
I'd compare it to Cassini-Huygens, more than Spirit and Opportunity.
"not much can go wrong with an airbag" -- not true. The mission designers for MER (Spirit/Opportunity) admitted that a wrong bounce or an initial collision with a sharp rock could spell doom.
I think the airbag relies considerably more on good luck. You can't really model collisions the way you can a descent on a cable. The latter is 100% engineered.
But as the video shows, the MERs did have a similar system - after the parachute phase, the rover in its cocoon was lowered from the aeroshell at the end of a tether. Then, the retrorockets in the aeroshell fired, bringing the whole thing to a full stop some tens of meters above the ground. After that, the airbags inflated, the tether was severed, and the rover dropped the rest of the way to the surface.
One of the reasons for the skycrane is that it minimizes the area of the Martian surface that will be disturbed by the landing rockets. The skycrane approach moves less surface soil around the landing site (decreasing travel distance to pristine subjects for geological study) and deposits less rocket exhaust (which complicates chemical and biological studies).
NASA has an excellent and free ebook, _When Biospheres Collide_ [1], which goes into great detail about the problems posed by biological contamination (both forward and back). The book dedicates a long chapter to the great lengths taken with the Viking landers to avoid contamination problems. Having read it, I'm not at all surprised NASA is giving alternatives a try.
The Viking program was resigned to regarding the surface soil as untrustworthy for at least some (if not all) of their experiments. To get around this, the landers' robotic arms dug a few inches into the Martian surface to reach uncontaminated soil.
As for driving, the other commenters have it right: Curiosity isn't particularly fast and it's nice to start the scientific experiments sooner, rather than later. Moreover, driving around consumes limited electrical power which has to be shared with all the other systems on board, including transmitters and scientific instruments. Every minute spent moving from place to place is a minute of reduced data collection.
They drive at mm/s speeds. Moving over a few meters because the soil is stirred up is a huge waste of energy and time, not to mention the risk of failure.
Actually, 30 m/hour is 8 mm/sec. But that's a bad choice of units, because seconds is not a good mission time scale.
The point is that, within a couple of hours, you can drive beyond the contamination radius. And, over the span of the mission, you most certainly will.
Sorry, I wasn't able to figure out where I first heard about the contamination concern with MSL. I think I may have heard it in a video interview, which makes it nigh unsearchable. That said, a search of nasa.gov reveals several mentions of the use of a skycrane to reduce rocket exhaust contamination in relation to other missions, particularly ExoMars.
