Come back when you have a counter-argument that isn't "It can't be done because we've never done it that way." If rocket people thought like that, there wouldn't be any rockets, because among the things said to be crazy were liquid fuel, lighweight turbopumps, swiveling engines, staging, vacuum operation, liquid hydrogen handling, and monocoque tanks, while the list of things that were thought to be essential were tail fins, launch rails, streamlined boat-tails, pointy noses, and parachutes. The 400 foot launch tower made out of pretty red trusswork: also not essential. It's a flying machine. It can fly on its own - unless your thinking is stuck in a tiny little box that dates back to the 1940's.
The Merlin 1D has a thrust to weight ratio of about 160:1. That means that if you wanted, you could have a completely seperate set of engines that are used only for 30 seconds, and you'd only pay a 1% mass penalty. It might end up lighter than adding elaborate tail-end landing legs and the fuel burn to transition from a horizontal return flight, bleed off the airspeed to zero, and gently touch down (a horizontal landing can kiss a runway at a hundred miles an hour, since it's not going to tip over).
And speaking of hover, pressure fed engines are even lighter than pump driven engines, and many months ago I thought about a demonstrator that was like a ball with thrusters aimed all around, so that it could rotate freely in the air on any axis while always having enough thrusters aimed vertically to maintain the hover. You can't do that with jet engines except by redirecting the exhaust because they don't have a thrust-to-weight ratio that allows for carrying many extras, and their throttle response is too slow. The SDI program actually had something almost similar, a missile that hovered horizontally and able to translate in any direction while remaining pointed at a fixed target. It could also, of course, pivot in any direction too, while hovering. Rockets don't have any problem doing that.